Practical Ways to Achieve Targets in Diabetes Care: Barbara Davis Center Keystone Conference 2013

July 18-21, 2013; Keystone, CO Full Report - Draft

Executive Highlights

In this report, we provide our full coverage of the compelling 2013 Barbara Davis Center Keystone Conference: Practical Ways to Achieve Targets in Diabetes Care. Conference director Dr. Satish Garg organized a four days of outstanding learning from some of the brightest minds in diabetes care, including Drs. Richard Bergenstal, Hans DeVries, Irl Hirsch, Fran Kaufman, Moshé Phillip, Robert Ratner, Robert Rizza, and Jay Skyler, just to name a few! Notably, this year included both pediatric and adult diabetes topics, a clear mark of how the conference has expanded; previously, Keystone alternated between the two topics and had a shorter meeting! To be sure, attendees always got their money’s worth and now, this could not be truer. In exemplary Keystone fashion, the 2014 agenda is already fully set – see the Barbara Davis Center’s website ( to see the speakers.

We counted 57 talks in the brochure (not including introductory remarks or industry sponsored dinners), with a fairly even balance between drugs, devices, and additional topics (treatment targets, healthcare system structure, reimbursement, etc.). Keystone’s strategic placement – just a few weeks after ADA –ensured there was plenty of debate and additional perspective on the hottest topics we learned about in Chicago in June. Sincere kudos to the Keystone organizers for making nearly all the speakers’ slides available online at– talk titles highlighted in blue are hyperlinked to slides, and you can click the red dates at the top to access each day of the conference. Editor’s note – next year, 51 talks are planned, about 10% fewer, though we imagine this may increase over the coming year.

This report organizes our comprehensive coverage into three topic areas: 1) Diabetes Technology; 2) Diabetes Drugs; and 3) Treatment Targets, Trends, and Additional Topics. Immediately below, you will find themes we’ve identified in each topic area:

  • a focus on the artificial pancreas;
  • the role of SMBG in type 2 diabetes;
  • enthusiasm for sulfonylureas;
  • discussion of cardiovascular risk and outcomes trials;
  • the role of public policy; and
  • the impact of health care reform; and 6) cost pressures.

As the conference continues to grow larger, there was talk of moving it to Denver, though a popular vote amongst attendees ensured that it will stay in Keystone through at least 2018. As noted, the brochure for Keystone 2014 is already posted (!), and boy are there some provocative titles – Dr. Philip Home: “Role of biosimilars in diabetes care” and Dr. Irl Hirsch: “How to manage T1DM Well in the Cheapest Way Possible.” You can view the full brochure at We’re already marking our calendars for July 17-20, 2014!

Table of Contents 

Keystone Themes


  •  The Artificial Pancreas (AP) was a key area of debate at Keystone 2013, especially regarding Medtronic’s ASPIRE in-home study, the use of glucagon, and liability concerns with the AP. On the heels of ADA 2013, the outpatient results of the MiniMed 530G were featured in talks by Dr. Richard Bergenstal (first author on the study’s NEJM publication) and Dr. Jay Skyler. Notably, Dr. Skyler (a member of Dexcom’s board of directors) was quite critical of the study during the Dexcom dinner, questioning whether basal insulin suspension was an effective strategy, or whether patients were merely having a study effect and consuming glucose when they were alerted to a low. However, he was very positive on the artificial pancreas overall and expressed enthusiasm for the use of glucagon, particularly in Dr. Ed Damiano and Steven Russell’s bionic pancreas (see our recent personal experience with the bionic pancreas at  or read about our visit to their camp study at Dr. Irl Hirsch’s outstanding talk on the “Steps to a Perfect Closed Loop” was similarly pro-glucagon, noting that its use allows AP systems to be more aggressive with insulin while simultaneously preventing exercise-induced hypoglycemia. Also on the glucagon front, Dr. Hans DeVries debuted a prototype of a dual-chamber insulin-glucagon pump to be used in the upcoming APPEL 4 study. That said, not all speakers spoke as positively on the hormone – Drs. Bergenstal and Moshe Phillip expressed greater enthusiasm for the near-term promise of insulin-only systems. Both speakers favored more of an incremental approach to the AP and cited the positive impact that insulin-only systems will have on reducing hypoglycemia.
    • Liability concerns related to the AP were an interesting topic of discussion as well – Dr. Hans DeVries drew the audience’s attention to the legal challenges of bringing a closed-loop system to market in the US. Dr. Robert Ratner also expressed his concern, asking, “What happens with the first car accident when the pump does not adjust to CGM reading?” Dr. Skyler’s answer during the Dexcom dinner harkened back to similar comments made during the first days of SMBG (when some doctors – and Dr. Skyler named names! – did not want to share the technology with patients). We thought his rejoinders were worthy of a standing ovation and were very heartened by this part of the meeting, for the same of patients. .
  • The role of SMBG in type 2 diabetes arose in discussion of the COMPASS study and in a presentation by Kaiser Permanente’s Dr. Andy Karter. Of note was Dr. Linong Ji’s discussion of the COMPASS trial, a Chinese trial examining the role of SMBG in people with type 2 diabetes. After three months, A1c dropped by an impressive 1.5% from a baseline of 8.5%, a change that was sustained at the six-month visit. Notably, patients who had a shorter duration of insulin use (<eight months) had a whopping 2.8% drop in A1c levels (from 10.3% to 7.5%), the largest of all the groups in the study. In line with previous successful studies in this area, the COMPASS trial used structured testing via the Accu-Chek 360 view tool – nice to see this being used quite broadly. Dr. Karter also emphasized use of structured testing in his talk on SMBG in type 2 diabetes – he hoped for more data on this front, as well as further study of PCPs’ willingness/skills to use SMBG to tailor medications. Perhaps most importantly, he called for the use of real-world, pragmatic trials of SMBG in type 2 diabetes, citing the flawed nature of RCTs to adequately establish a benefit. We certainly agree, as RCTs are often quite a far cry from real life and “control” groups do better than they otherwise would. Overall, it was encouraging to hear the discussion of SMBG in type 2 diabetes (especially given the current reimbursement landscape), a topic that was a focus area at Keystone 2012 as well – to read our coverage of the PRISMA Study presented last year, see page 52 of our report at



  • To our surprise, multiple panelists and speakers spoke positively on use of sulfonylureas. Dr. Robert Rizza said he uses sulfonylureas “all the time” at the Mayo Clinic, noting that the data linking them to decreased beta cell function is “little, if any.” We would like to learn more about which SFUs are used as we imagine this is more controlled than elsewhere and it would be great to get the Mayo education and experience out to a broader group. He said its “very hard to impune” them, despite the hypoglycemia and weight gain they cause; all speakers acknowledged these disadvantages, though did not seem concerned about them, except in the elderly. We were surprised about this especially because hypoglycemia and weight gain are significant issues in all patients, no matter what the age, race, socioeconomic status, etc. Dr. Satish Garg felt that sulfonylureas “have a very important role,” to the point where he uses them “a lot more than DPP-4s.” Both Dr. Garg and Dr. Rizza felt that the class has been “trashed”, partly because of pharmaceutical companies marketing new agents. That was interesting – we see, of course, better adherence with agents that have better side effect profiles. We wonder if doctors were being slightly politically correct, given all the cost concerns associates with diabetes. Dr. Richard Bergenstal highlighted that Park Nicollet Clinic which IDC is part of has “plenty” of sulfonylurea use too, mostly for cost reasons. And in a power-packed Meet the Peers session, Drs. Hans DeVries, Stephanie Amiel, George Alberti, and Angelyn Bethel all brought up benefits of the class as well – no cardiovascular disease disadvantage relative to new drugs (that we know of – these studies, of course, have never been reported on), very low cost (<$5 per year for treatment in sub-Saharan Africa), and decades of long-term data. We were very surprised to hear the notable Dr. Stephanie Amiel say, “Every time I open up the newspaper, I see new problems with new drugs.” Recent headlines associated with pancreatitis and new incretins drugs, for example, have never been proven, and for Dr. Amiel to imply it was fact was very disappointing. .
    • We would like to better understand adherence to SFUs as well as other orals – certainly, their convenience for patients and providers relative to injectables is higher, though we would be interested to see how adherence varies with side effects. There’s little doubt that lower drug cost is a major factor, but we would look for more focus on costs associated with complications. Focus on A1c continues as the principal metric for good diabetes control (i.e., SFUs tend to do well when diabetes control is assessed in this way). We wonder how the views would differ if DPP-4 inhibitors were generic. In that vein, it would also be valuable to see more cost-effectiveness studies, both for SFUs as well as other agents. We have heard that lower-dose usage of SFUs is more common in certain centers, and it would be terrific to seer the cost/benefit tradeoff of this – from talks with KOLs, it seems lower doses used closer to diagnosis is a more optimal time to use this class, broadly speaking.
    • We were surprised by how much enthusiasm we heard for this class given the aforementioned weaknesses, e.g., association with hypoglycemia, weight gain, and duration – that SFUs work very well for a shorter period of time and then effectiveness trails off. It will be very instructive to see how the SFU glimepiride does in Dr. David Nathan’s GRADE study (n=5,000, with an estimated primary completion date in 2020!), which will compare four drug classes head-to-head when added to metformin: glimepiride (generic SFU), sitagliptin (Merck’s Januvia), liraglutide (Novo Nordisk’s Victoza), and insulin glargine (Sanofi’s Lantus).
  • Cardiovascular safety and outcomes trials came up again and again at Keystone. Dr. Angelyn Bethel presented an entire talk on the issue of CVD outcomes in patients with type 2 diabetes, comparing statistics before and after the FDA came out with its 2008 guidelines. She highlighted that the number of CVOTs are growing in number, size and complexity, though one benefit is that studies now have harmonized cardiovascular composites. Dr. Hertzel Gerstein delivered two talks on the cardiovascular safety front, one focused on lessons learned from ORIGIN and another on evidence-based therapies for preventing CV outcomes. In the latter, he touched upon the outcomes of recent trials demonstrating CV neutrality of insulin glargine (ORIGIN) and saxagliptin (SAVOR), and also noted that SFUs are not particularly harmful in a cardiovascular sense (that’s good, since weight gain and hypoglycemia is enough trauma for patients). Meanwhile, Dr. Jay Skyler shared his views on the “glucose paradox” – despite the use of intensive insulin management in ACCORD, ADVANCE, and VADT, all three major trials demonstrated CVD neutrality – in his view, the neutral finding could be attributed to the trials’ short duration, insufficient power, and/or inclusion of well-controlled patients or those with existing CV issues. Last, cardiovascular safety was also discussed in sessions aimed at drug topics more generally. In his update on GLP-1 agonists and DPP-4 inhibitors, Dr. Robert Ratner highlighted the huge number of CVOTs being conducted on incretins, which will result in approximately 350,000 patient-years of data. Despite the enormous wealth of safety information that these trials will provide, Dr. Ratner specifically expressed concerns that investigators are not asking/answering the right questions about incretins and CV risk.



  • There was plenty to hear on public policy at Keystone, and talks on the ADA’s efforts and the impact of healthcare reform stood out. Dr. Jack Lord (University of Miami, FL) gave a quite valuable and focused talk on the Affordable Care Act (ACA), a law he characterized as “not real healthcare reform.” Dr. Lord detailed the positive and negative impacts that the ACA will have on diabetes care, highlighting that the broader availability of insurance will come with greater pressure on costs and strained access to providers. His three recommendations for industry followed accordingly: rigorously assess cost-effectiveness of products; acquire new skills in medical economics; and use “data, data, data.” Speaking in the same session, Dr. Robert Ratner (ADA, Alexandria, VA) provided an excellent overview of the ADA’s broad swath of public policy initiatives – we were particularly glad to hear about the Association’s active involvement at the FDA, its moves to get representatives on important government committees, efforts to conduct targeted research based on unanswered questions from special conferences, and an initiative to bring CME sessions to medical schools, academic hospitals, nursing schools, and graduate programs free of charge. There’s little doubt that cost pressures will remain in the coming years, though these talks gave us hope that industry and non-profit organizations can help overcome some of the challenges of operating in this environment. 
  • Panelists and attendees repeatedly questioned what the future of diabetes care will look like in an increasingly cost-constrained environment. In an informal “meet the peers” session of adult educators, speakers agreed that the majority of their patients are on Medicare, which covers just two hours of education per year (after the initial 10 hours of training in the first year). While telemedicine was raised as a possible alternative, said one educator, “We probably lose a great deal of money from this telemedicine…I’m more pessimistic about it with the budget changes.” We are also aware of awareness difficulties – many patients on Medicare do not come close to using these 12 hours. Meanwhile, attendees and speakers cited pay- for-performance policies in their reasoning for everything from whether or not to prescribe sulfonylureas to if an antibody negative child with type 1 diabetes should be screened for MODY. On a positive note, some had clearly accepted and were even embracing the need to focus more on cost-effectiveness in their clinical practices (e.g., instead of sending a vomiting type 1 patient to the ER, some pediatric providers favored the more cost-effective approach of prescribing the anti-nausea medication Zofran). Many also seemed to be fans of shared medical appointments, though there were lots of questions surrounding how to best conduct them. Other providers, however, were obviously frustrated with the restrictions reimbursement policies are imposing on the care they provide. Some working in independent education practices shared fears that their education centers could close due to tightening budgets. There will of course be winners and losers as costs tighten, though our hope is that the right things aren’t cut: diabetes prevention, education, and new therapies and technologies with better efficacy, improved side-effect profiles, and reduced burden on patients and providers.

Diabetes Technology

Opening Remarks

Introductory Remarks

Satish Garg, MD (University of Colorado, Aurora, CO)

Dr. Satish Garg, the conference’s co-planner with Dr. Jay Skyler, opened Keystone 2013 with introductory remarks. He noted that 38 US states (!) and nine countries are represented at this year’s conference, including several delegates making the long journey all the way from China – wow! Dr. Garg briefly honored the late Dr. George Eisenbarth, highlighting his greatness and the degree to which he will be missed – “for a very long time.” He humorously covered the conference’s dress code, explaining that “since you are in Colorado, at the most, business casual. Casual is preferred. But do wear something [Laughter].” Dr. Garg also asked attendees for honest feedback on the meeting, especially thoughts on whether to move the location to Denver in 2015 (depressingly, the number of attendees is getting too large for Keystone!). Last, he lamented one of the major changes in the schedule: Dr. Judith Fradkin’s (NIDDK, Bethesda, MD) inability to travel to Keystone due to government sequestration. Dr. Garg even offered to pay her travel costs, but government policy prevented this from being an option. What a loss this is and misuse of funds, given how much the government is spending unnecessarily on diabetes complications. We just saw Dr. Fradkin present on the first day of ADA 2013, and were disconcerted to hear her talk on “The NIH in the Era of Budget Cuts.” See the complete commentary on pages 215-216 at


Plenary: Closed Loop Options

Steps to a Perfect Closed Loop

Irl Hirsch, MD (University of Washington School of Medicine, Seattle, WA)

Dr. Irl Hirsch provided an outstanding review of what he thinks are the steps to a perfect closed-loop system: 1) measuring glucose at home accurately; 2) CGM accuracy following calibration; and 3) selection of an appropriate controller. We were particularly struck by how much his talk referenced payers and the challenging reimbursement environment – he was strongly negative on the impact competitive bidding will have on SMBG supplies, noting his concern that the policy will “cost us dearly in accuracy.” Turning to CGM, he showed data for Dexcom’s G4 Platinum (improvements since the Seven Plus and a very good ~90% severe hypoglycemia detection rate) and Medtronic’s Enlite (lower accuracy in Dr. Steven Russell’s independent study presented at ADA 2013 vs. the company-sponsored study from ADA 2012). Last, Dr. Hirsch addressed control algorithms, driving home the message that postprandial control is a serious challenge no matter which algorithm approach is used. He proposed several solutions, and seemed most positive on the use of a bi-hormonal, insulin-glucagon closed loop. In concluding, Dr. Hirsch cautioned that “we have to do better with closed loop than we do with our current technology,” as the current reimbursement environment will likely not support anything less.

  • Step 1: “If you’re doing closed loop and calibrating a sensor, you must measure glucose at home in an accurate manner.” Dr. Hirsch believes that current technology allows us to do this well in 2013. “However,” he noted, “as we move towards competitive bidding, first with Medicare patients and later with non-Medicare patients, the costs savings for strips will cost us dearly in accuracy, potentially sabotaging our ability to utilize the AP at Step 1 [Applause]…The potential infiltration of competitive bidding for Medicare patients for all patients using SMBG could potentially have a grave negative impact on the commercial viability of closed loop.”
    • Dr. Hirsch expressed particular concern over the accuracy of offshore meters. He showed encouraging accuracy data for Bayer’s Contour Next Link, highlighting its “very good data” in the hypoglycemia range (“right with YSI”). Dr. Hirsch’s worry creeped in, however, as he covered an accuracy comparison study from Brazg et al.,  JDST 2013. The study compared the accuracy of four offshore meters (Advocate RC, Embrace, True Balance, Element – we hadn’t heard of any of these!) vs. the Roche Accu-Chek Aviva Plus. Three sample lots from each meter were tested. All tested lots for the Aviva Plus met the 2003 and 2013 ISO standard; conversely, three of the 12 offshore meter lots did not make the 2003 standard, and zero of the 12 met the 2013 standard. Said Dr. Hirsch, “Would you want to calibrate your current CGM device with one of these meters? Would you want to make a decision on how much insulin to give based on one of these crappy meters?”
  • Step 2: Accuracy of CGM after calibration from fingerstick glucose. Dr. Hirsch noted that the industry agreed upon goal for the AP is a MARD <10%. He first showed data from industry-sponsored studies of the Dexcom G4 Platinum and Enlite, noting that the 13.2% MARD of the G4 is getting “very, very close” to <10%. Although the G4’s MARD in hypoglycemia is a bit under 20%, the severe hypoglycemia detection rate (percent of CGM values <70 mg/dl when YSI <55  mg/dl) is very encouraging, approaching 90%. Turning to the Enlite, the overall accuracy (from a Medtronic-sponsored study) was 13.6%, varying from a high of 15.9% on day one to 11.8% on day three.
    • Dr. Hirsch then posed a critical question: What if the CGM accuracy study is not sponsored by the company? Showing Dr. Steven Russell’s head-to-head-to-head data from ADA 2013 (see pages 62-63 at, Dr. Hirsch made two points clear: 1) in Dr. Russell’s study, Dexcom’s 10.8% MARD and 85% in the CEG Zone A were close to what was reported (13.2%) in Dexcom’s aforementioned pivotal study; and 2) Medtronic’s Enlite accuracy was 17.9%, higher than what the company originally reported. “The data are what they are,” he said.
    • To optimize CGM accuracy, it’s key to calibrate while the glucose is relatively stable (changing at <1 mg/dl/minute). Failing to do so can result in poor accuracy for an entire day. Dr. Hirsch emphasized that this is a “very important piece of advice” that the audience should take home to their patients.
    • Key questions posed by Dr. Hirsch: What is an acceptable high end MARD for a closed-loop system? Does the high-end MARD differ for a uni-hormonal and bi-hormonal (insulin-glucagon) system? What does CGM accuracy look like in a non-research, highly unsupervised setting?
  • Step 3: Selection of the appropriate controller/hormones. Dr. Hirsch quickly reviewed fuzzy logic, PID, and MPC algorithms, noting that all have the same problem: controlling postprandial hyperglycemia (at least with the current insulins). He listed several approaches to overcome this: faster insulin, heated insulin delivery, pramlintide, GLP-1 agonists, or simply more aggressive use of our current rapid-acting insulins. The problem with more aggressive use of insulin is hypoglycemia, which “brings us to bi-hormonal closed loop.”
    • “There are advantages to using glucagon. Over the last year or two, I have become much more sympathetic to this insulin-glucagon system if we are really going to do better.” Dr. Hirsch outlined two main advantages to adding glucagon to the AP. First, glucagon allows a closed-loop controller to be more aggressive and give more insulin, something that will help with controlling postprandial hyperglycemia. Second, a bi-hormonal system can prevent hypoglycemia from exercise, something insulin-only systems cannot typically do.
  • “In many of my adult patients, they can do just as well if not better than closed loop…if I cannot do better with a closed loop, I’m not sure it’s going to have much traction.” Dr. Hirsch believes this is especially the case with payers – in his view, it will be key to demonstrate superiority to standard of care. In our view, if the right patients are chosen, it shouldn’t be too difficult to show superiority with closed-loop control. What we wonder is how the data will look in very highly motivated, tech savvy, already well-controlled patients. We would guess closed loop will certainly benefit these patients on hypoglycemia, though we wonder if this will be enough for payers to reimburse it.


DREAM Project

Moshe Phillip, MD (Tel Aviv University, Petah Tikva, Israel)

Dr. Moshe Phillip presented early results from the first five patients participating in the second DREAM 4 pilot study of the MD-logic artificial pancreas system (MDLAP). As a reminder, the team’s admirable goal is to build a “glucositter” with the aim to prevent/reduce hypoglycemia while keeping patients in range. Dr. Phillip did not share specific statistics from the 3.5 weeks of outpatient overnight closed-loop control thus far, but it was clear from the graph that the overall interquartile range of glucose values shrunk; the study seems well on its way to the meeting the primary endpoint of reducing time spent <70 mg/dl. While not a primary focus of the team’s efforts, it was also encouraging to see that the median blood glucose of ~130 mg/dl improved over the ~145 mg/dl median blood glucose during the run-in period. The DREAM 4 second pilot study (n=30) has a similar trial design to the first pilot (n=44), whose interim results were also presented at ADA (for details on those results, please see page seven of our ADA 2013 Diabetes Technology Report at Participants will be on each arm (MDLAP or sensor augmented pump) for six weeks before crossing over to the other. The DREAM program has now included an impressive 167 patients for about 5,300 participant hours and about 450 nights at home.   

  • The second DREAM 4 overnight at home pilot study has a similar design to the first pilot. Participants (n=30) will be on either the MDLAP or SAP for six weeks, before crossing over to the other treatment arm for another six weeks. The primary endpoint is time spent with a blood glucose level <70 mg/dl.
  • Ten participants have completed the first 3.5 weeks of the study. Of these, five are using the MDLAP system and five are using the sensor augmented pump. Another twenty patients are in the run-in period.
  • Dr. Phillip did not discuss specific statistics for the study; however, from the overview graph he presented it was apparent that the overall interquartile range of glucose values was smaller in the MDLAP arm than the control.
  • Though users’ overall blood glucose levels are not a focus for the MDLAP-team (the focus is on hypoglycemia), we noted that median levels appeared to improve on the system. Among the five patients (representing 90 patient nights) on the MDLAP, the median blood glucose was ~130 mg/dl vs. ~145 mg/dl during the run-in period. The inter-quartile range of blood glucose levels was visibly broader and generally higher during the run-in period. 
    • One of the two example participants Dr. Phillip discussed had a mean glucose level on the MDLAP of 125±18 mg/dl over 18 nights vs. a mean of 176±43 mg/dl during the run-in period. The participant was 15 years old, and had a baseline A1c of 8.8% and BMI of 18.7 kg/m2.
    • The other participant had a mean glucose level of on the MDLAP of 125±14 mg/dl over 18 nights compared to his/her mean run-in level of 152±55 mg/dl. The 25 year-old participant had an A1c of 6.7% and a BMI of 26.5 kg/m2.
  • The DREAM program has now included 167 patients for about 5,300 participant hours and about 450 nights at home.
  • After participating in the second pilot study, one of the participants (an engineer in his profession, who has diabetes for more than 30 years – who participate in the study together with his son who has diabetes in the last 8 years) initially “refused” to return the MDLAP, according to Dr. Phillip. Dr. Phillip showed a video of the participant stating, “I don't care about all of the regulation – I will take [the MDLAP] like this.”
  • Results from the first DREAM 4 pilot study were presented by Dr. Revital Nimri (Schneider Children’s Medical Center of Israel, Petah Tikva, Israel) at ADA 2013 and has recently been published in Pediatric Diabetes. When considering the entire intent-to treat study population (n=44), MDLAP significantly decreased the time spent in hypoglycemia (blood glucose [BG] <70 mg/dl; 2.9% vs. 5.6%; p=0.02) and increased the time spent in euglycemia (BG 70-180 mg/dl; 81.5% vs. 73.6%; p=0.01). Similarly, the MDLAP system improved nighttime control when the analysis was restricted to adolescents and children. For more details on the results, please see page 7 of our ADA 2013 Report – Diabetes Technology (AP, CGM, SMBG, Pumps) at
  • Reflecting on his observations of patients as part of DREAM 4, Dr. Phillip expressed surprise at how “nighttime is not a sleepy time” for many patients. He explained that many participants were active, ate, and took insulin injections at night. Dr. Phillip suggested that this researchers and providers might want to have a conversation on the potential implications of this reality.


Future of the Artificial Pancreas at Home

Hans DeVries, MD (University of Amsterdam, Netherlands)

Dr. Hans DeVries began his talk by outlining four necessary developments for the artificial pancreas (AP): better continuous glucose monitoring (CGM), faster insulin, better pumps (although Dr. DeVries noted that this may not be the rate limiting step), and better algorithms. Turning to CGM, Dr. DeVries discussed Roche’s prototype sensor, which demonstrated a MARD of 9.2% vs. SMBG (pages 65-68 at – the first time a prototype sensor has reported a MARD under 10% (we agree the data is encouraging, though this product has a long way to go until it can be compared to already commercialized sensors). Discussing algorithms, Dr. DeVries noted that algorithms may soon be incorporated into pumps and sensors, thereby allowing manufacturers to compete and potentially reduce the price for patients. Dr. DeVries also briefly addressed APPEL 4, an upcoming bi-hormonal (insulin/glucagon) closed-loop study. Significantly, Dr. DeVries presented a prototype of the device to be used in the study: a dual-chamber pump that includes an integrated controller and an accelerometer. Looking ahead, Dr. DeVries wondered which of the dozens of companies developing an AP system will push forward with a commercializable device; yet, he also cautioned that some companies may be wary of advancing due to liability concerns (e.g., severe hypoglycemic events that could stem from insulin over-infusion). Dr. DeVries sees this as more of an issue here in the US. In closing, he emphasized that the AP must be made available to patient groups with sub-optimally controlled diabetes, as they stand to benefit the most from this technology.

  • Dr. Hans DeVries introduced a new prototype closed-loop system that integrates two pumps (insulin and glucagon), an accelerometer, and a controller – aside from Tandem’s artist rendering of a two-chamber pump (presented at JP Morgan earlier this year), this represents the first picture of a bi-hormonal insulin-glucagon pump we can recall seeing. The rectangular prototype weighs 280 grams (0.6 lbs) and is about the length of a pen and thickness of 1.5 decks of playing cards. It was pictured with two infusion sets emerging from one side.
    • The device will be used in the upcoming APPEL-4 bi-hormonal closed-loop study in the EU (two Medtronic sensors, a heart rate belt, and telemonitoring). While Dr. DeVries noted that the user interface still needs to be improved, the system is a drastic upgrade from the physical system used in the APPEL 3 study – that one was as clunky as they come and needed to be carried in a backpack! APPEL 4 participants will be monitored by automated technical failure detection. Dr. DeVries noted that 24/7 surveillance of participants becomes challenging as the studies get longer.
  • Dr. DeVries believes that bringing the closed loop to market in the US will be complicated, in part because of liability and legal issues. He explained that there is concern over liability for severe hypoglycemia, which would occur if a system over-doses insulin. While some believe that just one hypoglycemic event is too much, Dr. DeVries disagreed – in his view, clinical trials must make it clear that hypoglycemia improves with use of the closed loop, not that it avoids all events entirely. However, he mentioned that this is probably not a realistic view for the US. We think the liability question is a particularly fascinating topic and we look forward to better understanding how companies and the FDA are thinking about this. 
  • Dr. DeVries presented a list of eight companies he believes will be the big players for the closed loop system: Medtronic, J&J, Roche, BD, Dexcom, Tandem, Insulet, and Abbott. He believes that there will likely be only about five systems that actually make it to market. Dr. DeVries expressed his hope that these companies will make decisive moves about which research and algorithms to support, hopefully increasing competition and pushing prices down.
  • Dr. DeVries emphasized that it is important to look at how long a CGM is off target, not just the overall MARD score. He highlighted the Cambridge group’s findings that previous-gen CGMs could be off true blood glucose by 40% for more than an hour. When CGMs read off target for such long periods of time, it makes it challenging to develop aggressive algorithms. We would note that newer sensors like the G4 Platinum have improved considerably, though his point to look beyond MARD is a very good one.
  • Dosing insulin 15 minutes before a meal produces the best glycemic control vs. dosing at the start of the meal or 30 minutes before the meal. That said, we think it’s difficult for many patients to do this in practice, especially when eating out (i.e., it’s hard to count carbs before the plate gets to the table!). Meal announcement is a particularly interesting aspect of the AP, and one that needs a lot more research and refinement in our view. We particularly like Dr. Ed Damiano’s strategy, which aims to take the numbers and guessing out of pre-meal priming boluses. Rather, his team’s algorithm simply asks the user whether a meal is smaller than, equal to, or larger than typical. The algorithm doses accordingly, and then adapts over time based on its historical performance (i.e., giving more or less insulin up front the next time). 
  • Dr. Devries outlined two areas to improve insulin pumps: 1) improved occlusion alerts and 2) more precise insulin delivery. One study examined how long it took pumps to alert users to occlusion; the best pump reported after 27 minutes, while the worst pump did not report within 24 hours. In the latter case, the pressure did not build up enough to trigger the alarm, a major problem when thinking about closing the loop. (That said, we think advanced algorithms should be able to detect when the pump is dosing insulin but glucose is not falling). Additionally, there is room to make insulin dosing more precise and have doses in smaller increments, though Dr. DeVries is not completely sure that more precise dosing would add any benefit to treatment. We agree with his skepticism, and in our view, infusion sets represent a much more key area for improvement.
  • It would be a “bold statement” to say that there are no major differences in AP algorithms – more head to head studies are necessary. Dr. DeVries presented data from the CAT study, the first head to head study directly comparing two closed-loop algorithms (Cambridge vs. iAP). Dr. DeVries noted that there was no difference between the two – both showed a reduction in the time spent in hypoglycemia from 6% (open loop) to 2% (closed loop). For complete coverage of the trial, see page four of our report at
  • In reference to hardware problems, Dr. DeVries stated that the “only real solution is to start from scratch and make a new integrated device.” He remarked that he was pleased that companies are willing to have open communication between pumps and CGM, but communication can still fail. He noted that many companies are working toward integrating devices, specifically mentioning Medtronic and Drs. Ed Damiano and Steven Russell in Boston. He commented, “it’s not very sexy” work, but it is necessary.


ASPIRE In-Home Study Results

Richard Bergenstal, MD (International Diabetes Center at Park Nicollet, Minneapolis, MN)

Dr. Richard Bergenstal provided a summary of the ASPIRE in-home trial, originally presented in a late breaking poster at ADA 2013 and simultaneously published in the NEJM (Bergenstal et al., 2013) – see page 90 of our ADA 2013 report at The results showed a 38% reduction in mean area under the curve (magnitude plus duration) of nocturnal hypoglycemia events and a 32% reduction in nocturnal hypoglycemic events. The latter was a surprise since the device is not predictively suspending; Dr. Bergenstal attributed it to the view that “hypoglycemia begets hypoglycemia,” perhaps suggesting that less AUC for hypoglycemia could restore patients’ awareness and help prevent future events from happening. Most notably from a safety perspective, these strong efficacy results occurred without a risk for DKA or an increase in A1c levels (FDA’s biggest worries). New to us in Dr. Bergenstal’s presentation was his commentary on Dr. Bruce Bode’s ADA 2013 poster on the ASPIRE two-week run-in phase (where patients had to show episodes of nocturnal hypoglycemia to be included in the study). The analysis attempted to predict risk for nocturnal hypoglycemia, finding two key explanatory factors: high glycemic variability (the biggest predictor) and a lower baseline A1c (<7%) – we hope to see more analyses like supporting the clinical importance of glycemic variability.

  • Dr. Bergenstal highlighted a recent JAMA commentary (Mitka et al., 2013) on technology hazards in medicine – very importantly for CGM and the AP, number one was alarm hazards, including alarm fatigue. We hope to see much more research on alarms in the coming years, as this will become a bigger and bigger issue as more individuals get on CGM, closed-loop technology, and remote monitoring becomes more popular. Dr. Bruce Buckingham has presented excellent data on how CGM alarms fail to wake people with diabetes up at night. To us, this speaks to the ever-greater importance of automated insulin delivery.
  • Using landmark trials, Dr. Bergenstal highlighted how technology has improved rates of severe hypoglycemia over the years. Patients in the DCCT had 62 episodes of severe hypoglycemia per 100 patient years, or one every 1.6 years. In the JDRF CGM trial, this declined to 20 episodes per 100 patient years, or one every five years. STAR 3 brought it down further to 13.5 per 100 patient years, or one every 7.5 years. For Dr. Bergenstal, the ideal would be zero episodes per 100 patient years with an A1c of ~6.5%. Those in the ASPIRE sensor-augmented pump arm mirrored STAR-3, with 13 episodes per 100 patient years. On the other hand, those in the threshold suspend arm of ASPIRE had zero severe hypoglycemia events and an ending A1c of 7.2% (“We’re getting closer to ideal…though we’re not there yet”).


Panel Discussion

Irl Hirsch, MD (University of Washington School of Medicine, Seattle, WA); Richard Bergenstal, MD (International Diabetes Center at Park Nicollet, Minneapolis, MN); Peter Chase, MD (Barbara Davis Center, Aurora, CO); J. Hans DeVries, MD, PhD (Academic Medical Center, Amsterdam, Netherlands); Moshe Phillip, MD (Tel Aviv University, Petah Tikva, Israel)

Q: Since glucagon is part of the pancreas, why is it not included in the artificial pancreas?

Dr. DeVries: There are several issues with glucagon. First, current preparations on the market are for an emergency. You must reconstitute them at the moment of need. That’s a problem, of course, because there is lots of stress. We need a long-term stable glucagon. In experiments so far, you must reconstitute glucagon every 24 hours. The other reason is more philosophical. If you are a believer in insulin only, the current patient only has insulin. So why would you need a device with more than insulin? Many algorithms show reduced hypoglycemia, and perhaps we don’t need a second hormone.

Dr. Hirsch: It depends on the goal – it’s complicated. If you are going from an A1c of 13 to 10, you don’t need glucagon. You need a single hormone, insulin, in the closed loop if that’s the goal. But if you are starting with an A1c of 7.5% and you are trying to get down to 6.5% and reduce hypoglycemia, that’s different. One could also make the argument that if you have an AP, what about amylin? For somatostatin, I could argue we don’t need that. But there are recent publications on using pramlintide to reduce postprandial spikes. The real issue with insulin is trying to put the brakes on hypoglycemia. If some of these physical issues with glucagon can get worked out, I think that this going to be required to getting the ideal treat-to-target-almost-euglycemia without hypoglycemia. I don’t think it’s as doable as it would seem. That’s what we’ve seen with the group in Seattle.

Q: I would ask Dr. Bergenstal to comment on this. Since ASPIRE just used low glucose suspend (LGS), and now we have results from a predictive LGS, maybe we don’t need glucagon?

Dr. Bergenstal: We do not need glucagon to make great strides or to take a great step forward in minimizing hypoglycemia. I would say, “Lets not wait for the final solution.” Glucagon may be necessary for perfect control, but I am a believer in incremental steps.

Dr. Phillip: We have to deicide if we are going to wait for a perfect balance, and I absolutely agree we should move forward in incremental steps.

Dr. Hertzel Gerstein (McMaster University, Hamilton, Ontario, Canada): This was a wonderful session. There was a comment at the end about the types of patients in these studies. It’s very true. The severe hypoglycemia data from the DCCT was not just people on pumps; there is lots of it on MDI. I come from a jurisdiction where anyone with type 1 diabetes can get a free pump. But about a third are not interested in getting a pump. With all these CGM technologies getting better, there are approaches to mitigating hypoglycemia. What about something else that could be used for people not on a pump? Could you make it a lower tech thing? There is a real need for that.

Dr. DeVries: It’s an extremely powerful slide from Rich Bergenstal showing no severe hypoglycemia anymore in the ASPIRE trial. There has indeed been progress over time in reducing hypoglycemia. But the problem of severe hypoglycemia is still with us, and I’m sure that there are many populations where we don’t have the epidemiology. We should not be deluded into thinking it doesn’t exist.

Dr. Hirsch: What I think is the lowest tech way to prevent severe hypoglycemia, and that’s for adults, is to get a spouse! [Laughter] You’re laughing, but I’m very serious about this. I have had several patients that I’ve literally written prescriptions for. I tell them – I don’t think the insurers will cover this, but I’m going to write it anyways. Some patients put them on their refrigerators. We saw the reduction in severe hypoglycemia in ASPIRE with a three-month study. At least for me, it’s very difficult for us to make major conclusions from three months. In all of these studies, I would like to know for nocturnal hypos, was a patient alone at home or with someone? That to me is the most key issue with severe hypoglycemia.

Q: Does that work for husbands and wives?

Dr. Hirsch: I have a patient that wants a CGM so bad, but she couldn’t get it. Her boyfriend has Microsoft insurance. She did a common marriage, just so she could get the Microsoft insurance. That’s what they’re doing to get CGM in Washington state.

Q: I wondered a couple things about the practical use with the pump or AP. These technologies would be a lot smarter with a plan-ahead ability and some simple interface design. For example, you enter that ‘I am going to eat in 30 minutes’ or ‘I am going to exercise in 20 minutes.’ Could you anticipate any way for a simple plan-ahead feature in either the pump or the AP?  

Dr. Chase: We could teach temporary basal rates in training for people who go on a pump. I am so impressed when kids use these temporary basal rates. One fellow showed us a temporary basal he set before bed after exercising for hours – and he did not have any hypoglycemia. If we can get patients to think ahead and use these temporary basal rates, I think that would be fabulous.

Dr. Georgeanna Klingensmith (Barbara Davis Center, Aurora, CO): I would say that service dogs need to be continuously trained. If they become pets they are no good. On a light note, I totally agree on the spouse that doesn’t need to be a spouse. I have a young man and his parents are totally thrilled that he is moving into college with his girlfriend. They are totally relieved he won’t be alone at night. I tried to keep my mouth shut and my face neutral. On a more serious note, I would just take exception to getting someone from 13% to 10% - that you don’t need to worry about hypoglycemia. In the DCCT, those who had a higher A1c had a much higher hypoglycemia risk. When blood glucose is high, you give a lot of insulin.

Dr. Hirsch: That’s fair. But that was in the very different era of human insulin. We did things a lot differently back then. We don’t use those insulins now. The point was, I’m less concerned about using glucagon in that A1c of 13% population vs. an A1c of 7.5% population.

Q: Now we are getting more into this technology for preventing hypoglycemia, are we also seeing improvements in weight gain? Are these people gaining less weight than those in open loop?

Dr. DeVries: This is an intriguing question – when we extrapolate from the DCCT, for every lowering in A1c there is a decrease in body weight. This may not be the case in closed loop because there is no difference in the amount insulin patients receive, but only in the distribution of the insulin throughout the day. We may see a decrease in A1c without a decrease in weight. However, it is much too early to say.

Q: What are your thoughts on dogs who may be able to sense hypoglycemia?

Dr. Irl Hirsch: Some of my patients swear by the dogs. I would really like to see some study results that show if they work well or not.  

Dr. Chase: I know 12 families that use dogs, and patients who are looking to get dogs have to be careful because there are training systems that are charging unreasonably huge amounts of money for the dogs to be trained.  


Plenary: Glucose Monitoring

Is SMBG Needed for Type 2 diabetes?

Andy Karter, MD (Kaiser Permanente, Oakland, CA)

Kaiser Permanente’s Dr. Andy Karter provided a very academic look at SMBG for type 2 diabetes, emphasizing the many serious limitations in RCTs conducted to date – it was awesome to hear this. Dr. Karter highlighted that SMBG is a tool, not a therapy, a premise most studies to date have failed to embrace. Overall, he was quite positive on the use of SMBG in type 2 diabetes, but made it clear that the right patients, providers, healthcare setting, and education must be present. Dr. Karter concluded with a focus on key research needs in this area: 1) use of structured testing programs (STeP, PRISMA); 2) evaluations of PCPs’ willingness and skills to use self-monitoring to tailor medications; and 3) use of real-world, pragmatic trials (“the time is coming to an end for the classic randomized controlled trial for this type of study”).


Application of Current Guidelines in Type 2 Diabetes – COMPASS Study

Linong Ji, MD (Peking University People’s Hospital, Beijing, China)

Dr. Linong Ji presented preliminary results from the “Accu-Check missiOn: SMBG in Patients with diAbetes on inSulin Study” (COMPASS). In this two-phase, single-arm, prospective intervention study, Dr. Ji evaluated the benefits of self-monitoring of blood glucose (SMBG) on patients with type 2 diabetes. In the first phase, patients completed a questionnaire to determine baseline adherence to SMBG. Dr. Ji lamented that 65% of patients reported irregular blood glucose measurements, and 40% reported that they did not own a blood glucose meter at home and relied on hospital measurements. Phase two was an intervention study of ~800 patients with A1c levels >8%. Patients varied in the treatment of their diabetes, with the majority (63%) using premixed insulin (the preferred type of insulin therapy in China). Presenting preliminary results, Dr. Ji highlighted the ~2% drop in A1c levels after three months of increased SMBG. Patients with the shortest duration of diabetes (<5 years) and patients with the shortest duration of insulin use (<8 months) showed the largest drops in A1c levels (3% and ~2.8%, respectively). Dr. Ji remarked that analyses are still being performed regarding the effectiveness of the intervention, and his group is also planning on performing a patient follow-up. We were very excited to see a study that quantifies the benefits of blood glucose testing in type 2 diabetes, since SMBG is less emphasized in this population – especially early in the course of the disease where it seems to have the most benefit, according to these results. We hope further analysis will only serve to strengthen these results.

  • The questionnaire survey in phase one revealed that only ~2% of patients checked their blood glucose four times a day. Additionally, 65% of patients do not have a pattern of checking their blood glucose. The 40% of patients that do not own a meter at home relied on measurements taken at the hospital. Of the patients who took the survey, 51% used premixed insulin, 23% used basal insulin, and 7% used prandial insulin. The mean age of participants in phase one was 59 years, the average BMI was 25 kg/m2, and the average A1c was 8.5% (54% of participants had an A1c of above 8%). There were 1,353 male participants and 1,345 female participants.
  • The interventional second phase of the COMPASS study had primary endpoints of A1c, quality of life, and hypoglycemic events, although only A1c data were available for presentation. This phase consisted of 818 participants with A1c >8%. The mean age was 55 years, with 413 males and 405 females. The average BMI was 25 kg/m2, the average duration of diabetes was nine years, and the average duration of insulin treatment was 30 months. Patients in the study all used the Accu-Chek Integra meter A1c levels were evaluated at baseline, three months, and six months. Patients using pre-mixed insulin (63%) checked their blood glucose 2 times a day; those using basal insulin (15%) checked blood glucose 1 time a day; and those using basal-bolus insulin (15%) checked their blood glucose 3-4 times a day (protocol unspecified). Before each visit, patients did a 7-point by three day SMBG testing (via the Accu-Chek 360 view).
  • Across all durations of diabetes, A1c levels dropped ~2% at the three-month visit, and remained at that level at the six-month visit. Dr. Ji divided participants into four groups of diabetes duration (<~4.5 years, ~4.5-<~10 years, ~9-<~15 years, and >~15 years). Patients who had diabetes for <5 years had a 3% drop (from 10.5% to 7.5%) in A1c levels, the largest drop of all four groups.
    • Across all durations of insulin use, there was ~1.5% drop in A1c levels at the three-month visit, which persisted to the six-month visit. Dr. Ji divided participants into four durations of insulin use (<8 months, 8-<29 months, 30-<69 months, and >70 months). Patients who had a shorter duration of insulin use (<8 months) had a 2.8% drop in A1c levels (from 10.3% to 7.5%), the largest of all four groups.
    • Patients who have had diabetes longer or who have been on insulin treatment longer have a decreased probability of reaching A1c goals. Dr. Ji presented results that showed the percentage of patients meeting A1c goals decreased with an increase in diabetes duration. Similarly, the percentage decreased with the increased duration of insulin use. These results highlight the importance of early intervention for reaching and maintaining A1c goals.
    • Dr. Ji also noted that basal insulin use decreased and mixed insulin use increased with diabetes duration – not unexpectedly. While 23% of patients with a diabetes duration <5 years used basal insulin, only 10% with diabetes duration >15 years used basal insulin. Similarly, the percentage of patients using only basal insulin decreased as the duration of insulin use increased; the use of both basal and prandial insulin increased with a longer duration of insulin use (19% basal treatment with <8 months insulin use vs. 9% basal insulin use with >69 month use).


Role of CGM in Diabetes

Irl Hirsch, MD (University of Washington, Seattle, WA)

Dr. Irl Hirsch gave a broad overview of CGM, nearly identical to his excellent presentation in a hypoglycemia symposium at ADA 2013 (see pages 81-83 at and his presentation on CGM use in the T1D Exchange from ATTD 2013 (see pages 40-43 at He provided a view of the literature and devoted lots of time to the accuracy of Medtronic’s Enlite and Dexcom’s G4 Platinum. Overall, Dr. Hirsch concluded that CGM can improve hyperglycemic exposure and should have the potential to reduce the burden of hypoglycemia (the latter was said with caution – “I’m choosing those words very carefully”). He highlighted that the accuracy of the Enlite and G4 Platinum is less than perfect, especially in hypoglycemia and one day one, but it is improved from earlier generation devices. Like many aspects of diabetes therapy, Dr. Hirsch emphasized that patients need to be active participants in the use of CGM – “if you wear it, the A1c will come down.” Dr. Hirsch sees a future role for CGM in inpatients, type 2 diabetes (especially insulin requiring), gestational diabetes, neonates, and those with hypoglycemic disorders. He noted that in our current era of accountable care organizations, “a label from FDA is not enough.” The technology must demonstrate its value through improved outcomes. Last, he noted that “all of this technology is time intensive” – notably, he believes it’s not actually the technology itself, but the teaching, the downloading, and the follow-up required that adds most of the time and cost.

  • Are today’s CGM devices good enough to reduce the risk of severe hypoglycemia? Dr. Hirsch first quotes the current ADA and Endocrine Society guidelines, published in Diabetes Care. They note that CGMs represent an “evolving technology” that has made considerable progress, but accuracy of CGM is “poor” in the hypoglycemia range (60-73% of points in Zone A of the Clarke Error Grid). Dr. Hirsch countered that these guidelines were written based on data from earlier CGMs that are now being phased out.
    • Dr. Hirsch was positive on the Dexcom G4 Platinum, noting that it’s accuracy is approaching a MARD 10% (“We really need to get below 10%”). Still, he noted its lower accuracy in hypoglycemia, which is closer to 20%. As he did at ADA, however, Dr. Hirsch highlighted the G4 Platinum’s strong severe hypoglycemia detection rate, which approaches 90%.
    • Dr. Hirsch was a big fan of viewing CGM accuracy at various rates of change – he showed data from a Medtronic Enlite poster to illustrate how much rate of glucose change matters to accuracy.  Like his coverage of the G4 Platinum, Dr. Hirsch highlighted the Enlite’s higher inaccuracy in hypoglycemia (“It’s not horrible, but it’s not as good as we would like it”) and lower day one accuracy.


Role of Self-Monitoring of Blood Glucose and Continuous Glucose Monitoring – REACT Study

Richard Bergenstal, MD (Park Nicollet International Diabetes Center, St. Louis Park, MN)

Dr. Richard Bergenstal described the results of the REACT Study (originally presented at ADA 2013 as 170-OR), which randomized 104 patients with uncontrolled type 2 diabetes (A1c >7%) to either structured SMBG or real-time CGM (n=55 and 59, respectively). Patients were seen every two to four weeks for therapy adjustments, which were made by reviewing the Roche Accu-Check 360° View (SMBG) or the ambulatory glucose profile (CGM). After 16 weeks, patients in both the CGM and SMBG groups saw a significant decrease in A1c of ~0.8-1.1% from a baseline of 7.9-8.2% (p <0.001), though there was no significant between group difference. However, CGM did have a significant edge in reducing the percentage of readings in hypoglycemia. Dr. Bergenstal remarked that although glycemic control was similar in patients using both CGM and SMBG, the continuous nature of CGM helped make patients more aware of their patterns (i.e., with SMBG, it is much easier to miss a hypoglycemic or hyperglycemic event).

  • When looking at hypoglycemic data from the study, Dr. Bergenstal remarked that the values fell right around 70 mg/dl; however, there are big differences between different levels of hypoglycemia (e.g., a glucose level of 40 mg/dl is much different than one at 70 mg/dl). Dr. Bergenstal emphasized that we need to continue discussion surrounding blood glucose levels associated with hypoglycemia. For instance, he suggested that a severe hypoglycemic event is a separate category not related to its value at all, similar to how ketoacidosis is separated from its value in hyperglycemia. We agree with Dr. Bergenstal that not all hypoglycemia is equal, and we look forward to seeing how the guidelines shift as more and more people begin to understand the severity of extreme hypoglycemia.
  • For a full review of the results from this study from ADA 2013, please see page 37 of our report at


Panel Discussion

Irl Hirsch, MD (University of Washington, Seattle, WA); Andy Karter, MD (Kaiser Permanente, Oakland, CA); Richard Bergenstal, MD (International Diabetes Center at Park Nicollet, Minneapolis, MN); Angelyn Bethel, MD (University of Oxford, UK)

Q: I’m sensing a theme from this session – if you use it, things get better. If you prescribe SMBG to people who do it, A1cs come down. If you wear CGM, you do better. Certainly, if you improve delivery of data to providers and make it more interpretable, thing go better. How should we be moving our patients into this group of people that are knowledgeable and react intelligently? In this era of the 15-minute appointment, with most diabetes care in the PCP setting and difficulties in education reimbursement, how can we do better?

Dr. Bergenstal: In this day and age, it’s all about getting the patient involved. I come back to the medical home where you have some sort of support staff. The PCP, no matter how simple we make it, has no time to sit and digest the data. Someone else could pull the data and have it ready for the appointment. It should be all in the same fashion, so the PCP can get used to looking at it time and time again. Also doing it remotely, so you don’t have to check in person.

Dr. Karter: At Kaiser, we have concerns about getting patients on board to starting insulin. One thing that we’re doing is a program to start discussions way before they need to go on insulin. We’re starting the discussions early and introducing people to the idea. That’s something that we’re seeing over and over again – the importance of shared decision-making. Not having the provider tell the patient what to do, but having a joint discussion and coming up with solutions about moving forward. We’re doing a study right now on shared decision-making and adherence, and here is a really strong association. It’s very predictive. People are coming up with these fantastic new devices and wonderful fixes, but the pace is quite a bit faster than patients’ abilities. We have to be aware of health literacy and numeracy. That’s essentially the elephant in the room when you start talking about a real patient population. Can they handle this technology. It’s very important that providers are aware of that.

Dr. Hirsch: There are several elephants in the room. At Kaiser, IDC, and my clinic, we have an advantage. Here, we are talking about PCPs with 15-minute appointments. That’s not even what it’s like where I live. It’s under 10, and that’s very well documented. Even for endocrinologists, it’s very difficult to get 15 minutes for an appointment. You have the same amount of time in practice for a thyroid patient as a patient on a pump. That’s why we need a multidisciplinary team. I saw how Jay Skyler did that in the 1980s. We have tried to maintain that as we did in the DCCT. A couple weeks ago I was in Taiwan, and in the typical morning clinic, an endocrinologist will see 100 patients. Do that math. How do you do this technology with that, let alone an eight-minute PCP visit. Let alone all the other things we have to do that we get graded on for diabetes. To me, the greatest barrier is that until these technologies become more integrated with how we’re graded with ACOs, it’s going to be too big of a step to climb.

Q: I have more of a technical question for Dr. Bergenstal. Your last presentation was very nice, and it was nice data on whether or not blood glucose monitoring was effective in patients with type 2 diabetes. I have a question about real-time monitoring and SMBG. I was unsure of the scale on the y-axis for hypoglycemia. It looked like the values were only around 1-2%?

Dr. Bergenstal: That is correct. We are hopeful that these same results hold true in a higher risk group. We did as much as we could in a low-risk group.

Q: All this education that we got this week is wonderful. But how do we implement it in a 15-minute visit? Most of our patients are Medicare patients. They have no authorization for CGM, so we go back to self-monitoring. The problem that I have is non-compliance. Patients are not changing their diets or lifestyle. I work at the Indian reservations in Phoenix, AZ, which have high numbers of patients with diabetes. But they cannot get CGM, and my A1c is not going down; what do I do?

Dr. Bergenstal: The question is how do you get the patient engaged or involved? It’s shared decision making. It’s also about testing their blood sugar. How are you will displaying the data and collecting it and reviewing it?

Q: They have to go to diabetes education before they get the meter.

Dr. Bergenstal: But is the education how to use the meter?

Q: It’s a combo of diet, exercise, and how to manage glucose; they write it down and are supposed to bring it back for the next visit.

Dr. Hirsch: There are two issues. First, in most of the trials you’re seeing here, we’re very careful in terms of who we pick. There are eligibility criteria. In a study that Rich is involved with now, called FLAT-SUGAR, we were just talking about how you pick your patients is critical. But that may not work for your population. The point is that in these research trials, we have coordinators, educators, nurses, and PAs that spend inordinate amounts of time with patients that are not possible in real life. So when you look at this A1c data, it may not be replicable in the real world due to time constraints.

Point two has been an issue since the discovery of insulin: how do you take care of patients who are not interested in self-engagement with their own self-care? I see a few of these patients, but not as many as you do. At our county hospital in the University of Washington system, the majority of patients come in, but they are not interested in self-care. That’s a whole other topic that none of us has a really good grasp on.

Q: I am an associate professor, and I do community-based outreach education. I see these great clinical studies right now, but how do you start partnering more with community-based organizations? Those groups may be able to translate these results so that a community could get more education out of it – this may not be available in a short-term patient visit. I would also love to see some guidelines for the education – what HCPs think patients need and how community members could be effective in meeting these goals.

Dr. Karter: I agree.

Comment: I keep hearing all of these data, and I keep hearing the need for more pragmatic trials. I agree with that.

Dr. Karter: The new partner with NIH definitely agrees with you, and is definitely supporting this type of work with more patient research. Additionally, the RE-AIM [Reach Effectiveness Adoption Implementation Maintenance] framework is also in line with creating research that is more easily translated in a community setting.

Q: From your study, you did seven-point glucose monitoring in patients who were on oral agents. But most of the insurance companies will say check blood glucose one time per day. Or even every other day, or three times per week. Do we learn more by doing any one of these techniques?

Dr. Bergenstal: We don’t learn more if we don’t try it in a research setting. Having done that, you did not see a big difference in CGM data vs. a seven-point SMBG profile. There is a lot of discussion that if you have a lot of glucose data, and can see it, then you can give proper instructions about how to look at food and changes in blood glucose. We’re still learning.

Q: You showed that there were different rates of hypoglycemia whether you looked at <70, <60, or <50. If you look at serious hypoglycemia, the way DCCT defined it as assistance or passing out, what was the event rate?

Dr. Bergenstal: There wasn’t any – we didn’t have any severe hypoglycemia.

Q: For a long time, we’ve been thinking that people with type 2 diabetes are less prone to serious hypoglycemia. Is that what it looks like?

Dr. Bergenstal: We would need to take a group on basal-bolus therapy. If we’re really going to study that now, we need to see what populations are more prone. And we need to study that for a longer period of time.

Dr. Bethel: There are studies in the UK looking at different insulin regimens. They did CGM profiles and detected lots of asymptomatic severe hypoglycemia – that is not appreciated enough. We are just on the edge of understanding this.

Comment: I am so happy to hear discussions about what is going on with the patient. I have been working with a pilot program at the Barbara Davis Center to improve self-management with type 1 diabetes, especially in populations where that is difficult – for example, if the individual is a single parent or if they have a low-income. They were not excited about our cool technology that was going to save the day. So we went from a program that was cool-app-device driven to a program with a behavioral approach, such as motivational interviewing and problem solving skills training. It is extremely important to pay attention to where the patient is in everyday life and what approaches can be used. This goes way beyond education to make a difference. A lot of studies we have heard about have teams with psychosocial components, so maybe next year we could have a session discussion that.

Dr. Bergenstal: I have an additional comment on that topic. You did not see the data because I did not have it all done, but there were fewer medication changes than I thought there would be. A lot of the data showed behavioral changes, and I am not sure if this came through in the presentation. It is great to see such a drop in A1c levels without a large change in the medication regimen.

Comment: I have an answer for some of what Dr. Karter mentioned as follow-up questions. I wanted to call attention to a survey published in Diabetes Technology and Therapeutics as a follow-up to the STEP study. The majority of physicians are still using the 360 tool for patients with type 2 diabetes, and about half of their patients with type 1 diabetes.


Session: Adults: How Important is Glucose Variability in Diabetes

Glucose Variability – A New Standard: How to Measure It and Why it is Important

Hans DeVries, MD (University of Amsterdam, Netherlands)

Dr. Hans DeVries presented data supporting his preferred method of analyzing and describing glycemic variability, mean average glucose (MAG) – a single summation of all glucose change over time. Dr. DeVries began his presentation on the importance of glucose variability, remarking that measuring it is extremely helpful in predicting severe hypoglycemia and mortality in patients with type 1 diabetes. He highlighted that while standard deviation (SD) measures the dispersion of glucose variability, this measure is more open to noise than MAG. Another benefit to MAG, Dr. DeVries explained, is that many of the previous measures of glycemic variability have a high correlation amongst each other – for example, MAGE and SD have an R2 =0.88, which makes it difficult to view either measure independently. Dr. DeVries further asserted that MAG is a better measure of variability, mathematically speaking, and in one analysis, it was a better indicator of hospital mortality (p <0.001) than SD or area under the curve. He did note, however, that these results have not been replicated, so scientists have not been able to determine the best measure of variability. Still, Dr. DeVries believes that physicians should emphasize the importance of minimizing glycemic variability, especially because doing so can help reduce severe hypoglycemic events.

  • In line with his growing body of presentations on this topic, Dr. DeVries reviewed results from previous studies of glycemic variability. He highlighted that glycemic variability is not related to microvascular complications in patients with type 1 diabetes, and in patients with type 2 diabetes, is only related to oxidative stress when A1c levels are poorly controlled. He gave an overview of results from Monnier 2006 (JAMA) that originally suggested a strong correlation between glycemic variability and oxidative stress (R2 = 0.86), but turned out to be irreproducible later on by the same researchers.
  • Dr. DeVries built a case that MAG is the best way to measure glycemic variability. In the HEART2D study (Raz et al., Diabetes Care 2009), glycemic variability was visibly reduced when patients were treated with three doses of prandial insulin compared to one dose of basal insulin – a graph on Dr. DeVries’ slide made this point very clear, but interestingly, standard measures of glycemic variability did not capture the clear difference in glycemic variability (e.g., standard deviation, MAGE). However, MAG did, providing a very tangible example of why it could be a better way to measure glycemic variability.
  • Dr. DeVries mentioned a study similar to HEART2D, except 2-3x daily of exenatide would replace the prandial dose – he did not specify, but we assume this is Dr. Irl Hirsch’s FLAT-SUGAR study ( Identifier: NCT01524705). Dr. DeVries remarked that the trial is enrolling now (last updated on May 6, 2013 and currently recruiting participants). The trial has a primary completion date of July 2014 – as a reminder, it is a pilot study in 120 patients, and if it can successfully demonstrate differences in glycemic variability with comparable A1c levels in the two groups, there may be a subsequent full-scale clinical trial.


Workshop: CSII Treatment (Sponsored by Medtronic)

The Role of Technology in Diabetes Care

Francine Kaufman, MD (Children’s Hospital Los Angeles, Los Angeles, CA; Medtronic Diabetes, Northridge, CA)

Dr. Francine Kaufman led off the workshop by discussing the progress of continuous glucose monitors, threshold suspend (low glucose suspend outside the US), and the work toward an artificial pancreas. She emphasized that the artificial pancreas is not a complete “cure” – patients will still need to fill reservoirs, wear devices, and see physicians – but the goal would be to reduce the burden of diabetes and improve patient outcomes. Turning to threshold suspend, Dr. Kaufman noted that Medtronic is currently working with the FDA on requirements to get the MiniMed 530G threshold suspend pump approved in the US (the company recently received an approvable letter from the FDA; see our coverage at This is truly a regulatory saga for the diabetes device record books. On the bright side, we were encouraged to hear that Medtronic has been working with the FDA on the predictive low glucose management (PLGM) system (as of Medtronic’s F4Q13 call in late May, launch of the MiniMed 640G PLGM pump in Europe was expected by late November). In Q&A, Dr. Kaufman highlighted that by the end of 2013 or 2014, Medtronic will likely release in Europe the new dual sensor and infusion set (she revealed it’s unclear as to its commercial prospects). She also believes that Medtronic has moved a little slower than pump or sensor alone companies, since integrating sensors into pumps makes the device a “system.”

  • Dr. Kaufman highlighted the major studies performed outside the US on low glucose suspend (LGS, which is what the device is referred to outside the US) – all showed an improvement in hypoglycemia without an increase in hyperglycemia or A1c. The characteristics of LGS events were quite similar across all the studies – most happened every day or every other day, over 50% of events were less than five minutes, and only 10% of suspends lasted two hours (mostly at night). Dr. Kaufman also highlighted the striking LGS data presented by Dr. Trang Ly (Princess Margaret Hospital, Perth, Australia) at ADA 2013 – the number of severe hypoglycemia events dropped from six to zero following six months of LGS use is patients with hypoglycemia unawareness. To read our full coverage, see page 90 of our ADA 2013 full report at
  • The FDA’s main concern with low glucose suspend has been sensor inaccuracy resulting in an inappropriate suspend and DKA. However, Dr. Kaufman pointed to a study demonstrating that this is not a concern – even if insulin is shut off at a blood glucose level of 300 mg/dl for two hours, this would not be sufficient to put a patient into DKA. 
  • Dr. Kaufman reviewed the results of the ASPIRE In-Home study, presented in a late-breaking poster at ADA 2013 and simultaneously published in the NEJM (Bergenstal et al., 2013). For the full details, see page 90 of our ADA 2013 report at The results showed a 32% reduction in nocturnal hypoglycemic events and a 38% reduction in mean area under the curve (magnitude plus duration) of nocturnal hypoglycemia events. Notably, these benefits occurred without an increase in A1c levels.
    • Dr. Kaufman explained that the FDA required area under curve (AUC) reporting for the MiniMed 530G; the Medtronic team is currently in the process of sorting out how much of that was due to less time spent in hypoglycemia vs. a change in blood glucose levels. Results are also being analyzed by age and diabetes duration. As the FDA made clear in its artificial pancreas guidance, the MiniMed 530G is being reviewed as an artificial pancreas device system, and is thus being held to a high safety standard.
  • Medtronic is currently working with the FDA to answer questions necessary for the approval of the MiniMed 530G; Dr. Kaufman emphasized that Medtronic takes these requirements seriously and hopefully all of the agency’s concerns will be addressed soon. The latest timeline in Medtronic F4Q13 called for FDA approval by the end of 2013; we look forward to Medtronic’s upcoming earnings call for any updates on this front.
  • Studies of the predictive low glucose management (PLGM) pump in Europe are using a new pump platform and next-generation sensor (we believe this is Enlite 2, which was submitted for CE Mark approval as of May). Dr. Kaufman noted that if the MiniMed 530G continues to be delayed, then the US might soon be two CGM sensor generations behind Europe! Unreal.
    • Data on the PLGM system was presented at ADA by Dr. Thomas Danne – of the 16 patients who reached the hypoglycemic threshold for PLGM activation, PLGM was successfully activated in 15 of the experiments and prevented hypoglycemia (reference blood glucose ≤63 mg/dl) in 12 of the 15 experiments.
    • Here in the US, Medtronic has had a pre-IDE meeting for the PLGM pump. The FDA is requiring the rate of incidence of hypoglycemia (instead of AUC, which was used for LGS), which does raise the bar for efficacy. Still, the data on avoiding hypoglycemia events looks pretty good thus far, so we don’t expect this to be a barrier to approval.
  • Medtronic has plans to turn the data collected by CGM sensors into useful information. Eventually, Dr. Kaufman believes, separating out rates of change will allow physicians to give clinical suggestions to patients. Dr. Kaufman hopes to eventually have a cloud-based system where all data is sent seamlessly, and then relayed to patients’ or caregivers’ smartphones. [This describes Medtronic’s Connected Care system, which was on display in the ATTD 2013 Exhibit Hall. There is not current timeline for this product.] Dr. Kaufman articulately stated that the goal is to make diabetes management as simple for the patient as possible; perhaps so much so that diabetes begins to “look like pre-diabetes, which would be great.”
  • Dr. Kaufman emphasized that there is a racial variation in A1c levels that pump use does not get rid of; she highlighted this racial discrepancy is one area that needs work.

Questions and Answers

Q: You have talked about the pump, the sensor, and the meter, but what about the insulin? Are you going to work toward faster acting insulin?

A: Great question. We are not an insulin company, but we will partner to improve patient satisfaction and outcomes using our devices. We are interested, as expected, in increasing the speed at which insulin is absorbed.  So new infusion sets, different placement such as the intradermal space, heating elements or local compounds to increase absorption, as long as they don’t increase patient burden, are of interest to us.  Another idea, like you mentioned, is to look at insulin. As soon as an ultra-rapid insulin comes along, everyone is going to evaluate its use with pumps. Additionally, another idea is peritoneal insulin delivery. However, we have to look at whether they are more dangerous or more work – and the FDA… I can’t even imagine that road to approval. There are multiple elements that go into trying to build this system.

Q: I have a couple of questions about the output reports from sensors. Sometimes the patients’ reports [CareLink Personal] and our reports [CareLink Pro] do not match. If they do not match, patients tend to prefer our reports. Is there a movement toward more similarity between the reports? Also, I am very rarely able to get reports for kids because the program requires at least five days of continuous wear, which is hard. Is there any movement toward making this time period shorter?

A: I envision CareLink as some big Queen Mary ship floating in the sky – I cannot quite understand where all that data is. It is a little bit of a bottleneck for us because every technology we develop requires a new reporting method. All of that takes coding and has to go through FDA. We will likely put all of the information on the web, and HCPs can do with it what they like. People get upset when we take things away. We are afraid to stop reporting anything, so we will let you decide what you would like to use. Speaking to your second question, if there is not enough sensor wear, there is nothing to analyze. I think we are going to work to aggregate the information of a few weeks, but aggregating information that is very distant may not provide a very accurate picture with what is going on with a patient. Hopefully patients will be enthusiastic about wearing the sensor with the new changes we have made to it.

Q: Do you have a timeline for when the new sensor will come out?

A: I don’t have a timeline! I think one problem that may be unique to us is that when you develop both a sensor and a pump, everything is about the system. It would be easier to release a sensor that does not work with a pump. We have gotten ourselves into the mindset that everything has to work with everything else. I think in the future we will not look backward; if you want the new device, you have to buy the whole new system. There are lots of things coming. I invite you to come to my dinner tonight to hear about my new project going on in Kazakhstan.

Q: We know that it is good for kids to wear their sensors; however, they usually do not even if it is just sitting in their home. How do we encourage kids to wear their sensors?

A: Enlite and our generations of sensors that follow are very different than are present sensor. When people are allowed to get these new sensors, consistent sensor wear in pediatric patients increases significantly. This is more evidence that we have to reduce patient burden. That is what we are working on with design and accuracy. Dexcom is moving in the same direction. They are moving faster than we can because we link back to the pump.

By the end of this year or next, we will likely release in Europe the new dual sensor and infusion set.  One side is the infusion and one side is the transmitter. The sensor doesn’t actually care that the insulin is being delivered; the insulin does not interfere with signal. Another company is looking to put the sensor inside infusion set. But the goal is to reduce burden. The hope is that patients will wear the system more.


Optimizing Control with Therapy Management Software

Linda Burkett, MSN, RN, CDE (Clinical Education Specialist, Northridge, CA)

Ms. Linda Burkett provided a review of the information available from reports put out by CareLink web-based software. She reiterated the commonly accepted (and very important) wisdom – patients that look at their data do better. (Of course, they also need to know what TO DO with the data once they are looking at it, often not an easy task in practice.) She noted that the aim of CareLink is to provide a clearer picture of diabetes and help clinicians identify patterns. Ms. Burkett highlighted that about two and a half years ago, the therapy management dashboard and the episode summary were added into the program. These updates allow HCPs to observe how many basal rates patients are on, how much variation there is in basal rates, what carbohydrate ratio is being used, and what alert settings are enabled. Overlays of the data provide helpful information in determining if patients are experiencing low blood glucose levels at specific times of the day or after certain types of activity. Additionally, Ms. Burkett noted that CareLink provides important behavioral data, such as how often patients are testing their blood glucose and if patients are using their bolus wizard. Looking forward, she explained that future versions of CareLink will include more analytics of sensor data and more connectivity between patients and families – we think these are important and critical steps going forward (and certainly key for keeping up with Dexcom’s future plans with Share, Gen 5, Qualcomm, and SweetSpot).

  • Ms. Burkett emphasized that the key to diabetes management is empowering patients and shifting barriers to successes. Specifically, she encouraged HCPs to show CareLink reports to patients and ask them to find positive aspects in their diabetes management. Additionally, Ms. Burkett noted that HCPs should write on reports and send them home with patients to help them manage their care outside of the office. HCPs should encourage patients to connect diabetes self-care to personal goals.

Questions and Answers

Q: Can you see all of this data if patients are not wearing a sensor?

A: If patients are not wearing a sensor, you cannot see the episode summary and therapy management dashboard. Most of the data on the therapy management dashboard you can get from other places on the report – but there is some data you cannot. This is because these values are hard to get from only two or three or four blood sugar values a day.


Crack the Case

Kim Larson, MS, RD, BC-ADM (District Clinical Supervisor, Denver, CO)

Concluding the workshop, Ms. Kim Larson presented a new way to evaluate patient sensor data in CareLink using the “Yes! iCAN” method: “Yes!” (point out what the patient is doing correctly), Identify, Cause, ActioN. She explained that the current methods of evaluating diabetes management (1-2-3, Triage, and CareLink Progress) all have benefits, but there needs to be a way to standardize CareLink interpretation going forward. She identified two major points considered when creating the new methodology: 1) Staying positive is key. Whether the data is mostly good or bad, CareLink interpretation can take on a negative undertone, and patients may feel that their areas for improvement are disproportionately targeted; and 2) the new method must be flexible. The “Yes! iCAN” framework is intended to help HCPs determine the most important pieces of information quickly during patient visits. We were excited to hear that a standard method of CareLink interpretation is developing, since data interpretation is so challenging and often consumes too much of in-clinic time. We are also huge fans of the emphasis on staying positive, since diabetes management can be unpredictable and disheartening at times.

  • “Yes!” The first step is to identify what the patient is doing correctly. Ms. Larson emphasized that all patients do something well when managing their diabetes, and it is important to encourage patients to continue.
  • Identify. Does the patient struggle with hypo- or hyperglycemia? Or are most of their challenges behavioral? With the vast amount of data output by CareLink, it is key that HCPs sort through the information and seek patterns that help them identify real issues.
  • Cause. The HCP should then identify the cause of the previously identified challenges. These can be either concrete causes, or questions the HCP needs answered. Both of these, she noted, are key to providing good care for patients.
  • ActioN. Finally, Ms. Larson explained that the action HCPs decide to take can be either a clinical or behavioral adjustment, such as using the bolus wizard more frequently. She emphasized that only one or two changes should be made at each visit, allowing HCPs to understand which change is affecting outcomes (and of course, not overwhelming patients with too much to think about).


Workshop: Continuous Glucose Monitoring (Sponsored by Dexcom)

Dexcom Technology and Future

Jake Leach (Dexcom, San Diego, CA)

Mr. Jake Leach’s presentation focused on developments in CGM technology, and prominently featured the many advances with the G4 Platinum. The biocompatibility and flexibility of the new sensor reduce post-insertion wound response, leading to improved performance on day one, more stable calibration, and extended reliable duration of use. Mr. Leach discussed Dexcom’s many partnership projects, noting that the Animas Vibe is currently under FDA review – as of Dexcom’s 1Q13 call a potential launch could happen in 4Q13 or early 2014 at Animas’ discretion. Mr. Leach also addressed the Tandem t:slim integration partnership, which is poised for an FDA filing in the near future – per Dexcom’s 1Q13 call, filing was targeted before year-end. It’s impressive to see how fast that project has moved. Mr. Leach ended on the exciting topic of closed-loop systems, pointing to the large number of trials using the G4 Platinum, including an upcoming “control to range” study funded by the JDRF and the European Artificial Pancreas Consortium’s AP@home outpatient feasibility study.

Selected Questions and Answers

Q: How will calibration work for real world closed-loop studies?

A: In previous studies we used the OneTouch Ultra, but right now we’re in the process of looking into what meters to use in our studies moving forward. Roche’s Accu-Chek worked well, among other meters. The most important thing though is the proper cleaning of fingers, as it has been one of our biggest sources of error.

Q: I work in pediatrics, where probably the biggest threat to the technology is poor calibration technique and error or fabrication in entering the calibration results.

A: As we’ve made the sensor technology better, we start relying less and less on calibration. Previously, we would always trust the fingerstick over the sensor. Now we’re starting to trust the sensor as much as the fingerstick, and the smart sensor asks for more fingersticks if a calibration value doesn’t make sense.

Mr. Terry Gregg (CEO, Dexcom, San Diego, CA): When we talk about our goal to replace fingersticks, a big next move is to move to factory calibration. I can’t tell you if we will be successful, but based on past experience, we should be. The goal is to simplify the process and eliminate the variability, especially the human variability. For future devices, you may only have to do a post-insertion calibration and you’ll be set for the next week. It’s our guess that it’ll take us at least a hundred million dollars of investment to get to the point where we can have those types of sensors. We put about forty percent of our income into R&D, as compared to the ten percent most companies do. We’re sensor technologists and our pipeline is rich with things for the future. It’s hard to define exactly when these things will happen, but its probably around the five to seven year timeline we mentioned earlier.


Limitations and the Pathway Forward

Damon Tanton, MD (Florida Hospital Celebration Health, Celebration, FL)

Dr. Tanton began his talk by introducing the word “passionary,” a combination of “passionate” and “visionary,” to describe the attitude he feels is necessary when discussing continuous glucose monitoring (CGM). He started by surveying the present CGM landscape, noting that the technology is generally associated with motivated, insulin-dependent type 1 diabetes patients, and is usually prescribed by endocrinologists because of its perceived complexity. He predicted that those trends could reverse themselves in less than a decade, leading to more type 2 diabetes patients on CGM and shifting prescription responsibility to primary care physicians. Having personally tried monitors from Medtronic and Dexcom (“If you’re going to prescribe them, wear them” – this was awesome to hear), he opined that the G4 Platinum is the best device on the market. He specifically mentioned the G4 Platinum’s accuracy, which he characterized as on par with fingersticks and with less lag time than earlier devices. Looking forward, Dr. Tanton believes that CGM should, in the future, dethrone fingersticks as the “gold standard” of SMBG technology – we agree, though the cost must certainly come down to make this possible. He also advocated that CGM should be available to people with prediabetes in the future.

Selected Questions and Answers

Q: In your practice, with type 2 diabetes patients, are you using CGM in patients on oral agents along with the ones on insulin?

A: Usually no. I generally only use CGM in patients on insulin. In a few cases I use CGM in patients for other reasons, such as frequent hypoglycemia, but generally no.

Q: What are the age criteria for CGM?

A: I’m not a pediatrician so my minimum age is eighteen.

Audience Comment: I don’t seem to have any problem getting children to put on CGM. There are sometimes some specific insurance requirements, but I usually just do it and it’s fine.

Mr. Terry Gregg (CEO, Dexcom, San Diego, CA): We’re currently looking at a fall timeframe for getting a pediatric indication. [Editor’s Note: This is right on par with Dexcom’s 4Q12 call, where management guided for a 2H13 approval.]


Training and Device Demo

Keri Weindel, MS, RD, CDE (Dexcom, San Diego, CA)

Ms. Keri Weindel ran the hands-on segment of the workshop, allowing attendees to insert and operate the Dexcom G4 Platinum. She began by noting that Dexcom is unique in the CGM space in that it has an approved self-training program for new users. We think this is a huge positive for patients, since it reduces the hassle factor of starting on CGM. After discussing the updated G4 Platinum receiver, she walked the room through the sensor insertion procedure. Attendees could also demo the device’s alert system, making the room sound like (in the words of one audience member) a casino game floor. Attendees generally seemed impressed by the ease of insertion and lack of pain they experienced. The event ended with a discussion of the G4 Platinum’s contraindications, which include acetaminophen and MRI/CT scans and diathermy treatment. [Though we've seen recent data that this is less of a concern than it was with the Dexcom Seven Plus.]

Selected Questions and Answers

Q: Is there anyone here who has tried the Medtronic device, and if so, was there a difference between that and the G4?

Audience Comment: Yes, there was definitely a huge difference. With the Medtronic, I could feel it for a long time after insertion and after a day I couldn’t wait to take it off.

Audience Comment: I think that the easier insertion will improve patient adherence.


Patty Telgener, RN, MBA (Emerson Consultants, Excelsior, MN)

Ms. Patty Telgener provided a comprehensive, high-level overview of the current reimbursement landscape for diabetes care, focusing on CGM. She touched upon the hot topic of CMS’ new competitive bidding process (read our latest update [July 1] on competitive bidding at, noting that it currently applies to mail-order BGM supplies, and by January, might be extended to insulin pumps (negotiations are reportedly ongoing). She also discussed the “anti-switching rule” that prevents contract suppliers from influencing beneficiaries to switch to another brand. Turning to CGM, Ms. Telgener discussed reimbursement coding for professional and personal CGM, very similar to Dr. Claudia Graham’s talk at AACE 2013 (pages 56-58 of our report at In concluding, she highlighted the “next phase of CGM reimbursement,” including the expansion of coverage for type 2 patients, CGM coverage under pharmacy benefits, and Medicare coverage for personal CGM. [On the type 2 front, Dexcom’s 1Q13 call shared encouraging news: in the last year, 12 private payers (representing ~20 million covered lives) have amended their policies to provide coverage of CGM for insulin-using type 2 patients] Ms. Telgener acknowledged that, although coverage for CGM has continued to improve, most patients must still bear significant out-of-pocket costs for the technology. As we understand it, the out of pocket costs are related to how insurance plans are structured (e.g., employers and patients are choosing plans with higher out of pocket/deductibles to try to reduce premiums). While new clinical studies will continue to expand coverage decisions for CGM, out of pocket expenses will depend on the structure of the patient’s plan.  

Selected Questions and Answers

Q: Does Medicaid cover CGM?

A: It’s a state-run program, so what I say for one state would not apply in many others. Some Medicaid programs are covering it, some are not. It comes down to understanding your specific state.


Industry Sponsored Dinner (Sponsored by Dexcom)


Terry Gregg (CEO, Dexcom, San Diego, CA)

Mr. Terry Gregg provided introductory remarks to open the Dexcom dinner symposium, highlighting that “G4 Platinum is here today. It is highly accurate. It is simple to use. Those are the attributes that are most important to your patients to ensure compliance and usability.” To support his point, Mr. Gregg discussed a recent Kaiser study looking at utilization rates of CGM. Notably, over 90% of patients that started on the G4 Platinum have continued to use it – striking data that we hope Kaiser shares in more detail in a publication or at a conference. Concluding, Mr. Gregg stated, “We look to the future, and our goal is to constantly innovate and bring forth more accurate sensors, make them easier to use, and have patients use them more effectively and efficiently. The future is here today, and that’s why we’re here.”


Real-Time CGM: 2013 and Beyond

Jay Skyler, MD (University of Miami Miller School of Medicine, Miami, FL)

Dr. Jay Skyler, a member of Dexcom’s board of directors, led off the Dexcom dinner with enthusiasm, noting that this is his favorite meeting of the year. His data-driven presentation addressed five primary areas: 1) the JDRF CGM trial (“the single most important paper to come out about CGM”); 2) the historical relationship between A1c and severe hypoglycemia risk; 3) his critique of the recently reported ASPIRE in-home trial of Medtronic’s MiniMed 530G (since the median duration of suspend events was 11.9 minutes, it likely wasn’t low glucose suspend that reduced hypoglycemia; rather, it was patients responding to the alerts and consuming glucose); 4) comparative accuracy studies between CGMs on the market (the focus was on Dexcom’s superiority relative to Medtronic); and 5) some of the most exciting artificial pancreas studies (Dr. Ed Damiano’s Beacon Hill study and Dr. Boris Kovatchev’s outpatient study, both discussed at ADA 2013). Dr. Skyler closed by emphasizing his main message to the audience, “Go for accuracy. Manage your patients with good education. Download your patients’ data and use it.”

  • Concerning the recently reported ASPIRE in-home trial of the MiniMed 530G, Dr. Skyler referred to Dr. Irl Hirsch’s editorial that just appeared in DT&T – in that piece, Dr. Hirsch expressed his surprise that the median duration of suspension was only 11.9 minutes. He also acknowledged two limitations of the trial: a short duration (i.e., reduction of hypoglycemia over time is not known) and benefits in adolescents, young adults, or geriatrics cannot be extrapolated.
    • Was basal insulin suspension effective? Or were patients in ASPIRE in-home having a study effect and consuming glucose when they were alerted? Dr. Skyler highlighted the key implication of a median 11.9-minute cessation of insulin in ASPIRE in-home – 50% of cessations were less than 11.9 minutes. Given the short duration, he hypothesized that these patients were being alerted to lows, performing SMBG, treating their low with glucose, and resuming insulin delivery. Assuming basal insulin delivery of one-unit per hour, the 11.9-minute suspension equates to 0.2 units of insulin (“Do you really think that would be enough?). We thought this was an astute and persuasive point, though it is also largely in line with what Medtronic has found in other studies of LGS – the vast majority of suspends are less than ten minutes (and mostly during the day), while most two-hour suspends occur at night. Dr. Skyler did not specifically address full two-hour suspensions of insulin at night, which is what many believe is the real value of the LGS device. Broadly speaking, we think it’s extremely challenging to tease out whether the overall study effect of the device in ASPIRE was due to patients treating their lows (in the case of short suspends), the longer two-hour MiniMed 530G suspends, or some combination. 
    • Dr. Skyler questioned that value of suspending basal insulin delivery to correct hypoglycemia. He pulled up the 2013 ADA Standards of Care, which call for using glucose as the preferred treatment to correct hypoglycemia. Dr. Skyler also noted the long tail of insulin analogs – according to Dr. Lutz Heinemann, it “takes 2.5-4 hours until a considerable change in basal infusion rate leads to a new steady state level in the induced metabolic effect.” Said Dr. Skyler, “Fundamentally, patients were treating themselves better with glucose. It’s not the system. You need an accurate glucose monitor – detect hypoglycemia with CGM and treat it rather than suspending basal insulin delivery.”
  • Dr. Skyler shared recent independently conducted CGM accuracy studies – results consistently had the Dexcom G4 Platinum on top, with a MARD ~10%. Two studies were just recently published in the Journal of Diabetes Science & Technology (Pleus et al. and Freckmann et al). Pleus et al. found an overall MARD of 10.9% with the G4 Platinum and a precision absolute relative difference (PARD; putting two sensors on people and comparing the differences between paired CGM readings) of 7.3%. In the same issue, a companion paper by Freckmann et al. (from the same German group as Pleus) published a comparative CGM accuracy study, finding a MARD of 12.4% for the Abbott FreeStyle Navigator, 16.4% for the Medtronic Enlite, and 16.7% for the Dexcom Seven Plus. Notably, Roche’s new sensor (in another companion paper) came in at a MARD of 9.2%, though Dr. Skyler cautioned that it is still in the experimental stage (i.e., “made in a limited number by hand in a lab”; sort of an apples and oranges comparison since the Roche sensor is not manufactured at scale). We note that these Freckmann et al. results were very much in line with the third study Dr. Skyler covered – Dr. Steven Russell’s oral presentation at ADA 2013. In that head-to-head-to-head trial, the G4 Platinum had a MARD of 10.8%, compared to 12.3% for the Abbott Navigator and 17.9% for the Medtronic Enlite. See pages 62-63 at for complete coverage.
  • In his discussion on the closed loop, Dr. Skyler expressed enthusiasm for Dr. Ed Damiano and Steven Russell’s work: “To me, the Boston guys are at the top of their game…they are moving along rather rapidly.” Dr. Skyler highlighted some of the preliminary Beacon Hill results presented at the Joslin Symposium in May and ADA 2013, and we just saw the most updated results from 10 patients at Friends for Life a week and a half ago: a mean blood glucose of 130 mg/dl during closed loop (days four and five) vs. 151 mg/dl in open loop. Additionally, just 0.4% of the time spent was <60 mg/dl (days four and five) vs. 4.8% of the time during open loop. Dr. Skyler called the results “pretty remarkable,” and we know he doesn’t dole out that kind of praise lightly! As a reminder, the Boston team is currently conducting a 32-patient diabetes camp study this summer, meaning by this fall, 52 outpatient experiments will be completed!
    • Like Dr. Hirsch earlier in the day, Dr. Skyler emphasized that adding glucagon into the system allows the Boston group to aim lower. This was immediately apparent when he discussed the other most exciting closed-loop study from ADA – Dr. Boris Kovatchev’s outpatient study testing the DiAs smartphone platform, Dexcom G4 Platinum, and Tandem t:slim. The insulin-only system achieved a 36% reduction in the risk of hypoglycemia, though at a much higher average blood glucose of 161 mg/dl. Said Dr. Skyler, “It’s higher because they don’t have glucagon.” For further coverage of the study, see page 49 at
  • Dr. Jay Skyler gave an excellent review of the JDRF CGM trial, calling it “the single most important paper to come out about CGM.” He succinctly summarized many of the study’s key findings: 1) Sensor wear drives outcomes (“You can’t expect CGM to work unless you use it. Duh!”); 2) declines in A1c were sustained at 12 months (“If you actually use it, you continue to get beneficial effects”); 3) CGM works in people that are already well controlled, as well as those who need to improve.
    • Notably, Dr. Skyler believes that the STAR-3 study has been misinterpreted. The study compared pump plus CGM to MDI plus SMBG. “It’s apples and oranges,” he noted. “What they should have done, but allegedly due to lack of funding, is compare MDI plus CGM to a pump plus CGM, or pump plus SMBG versus pump plus CGM. If you put two things against two things, you misinterpret the results.” Conversely, the JDRF CGM trial was done “the right way.”
  • Like other speakers at Keystone, Dr. Skyler highlighted the improvements in severe hypoglycemia since the DCCT, moving from 60 events per 100 patient years in the DCCT (1986-1993) to 20 events per 100 patients year in the JDRF CGM control group (SMBG only; 2006-2007). Dr. Skyler attributed the dramatic improvement to the switch from regular/NPH/Ultralente in the DCCT to rapid-acting and basal insulin analogs in the JDRF CGM trial. Adding in the JDRF CGM intervention groups to the slide essentially eliminated the real risk of severe hypoglycemia by adding CGM – less than 10 events per 100 patient years.


Dexcom Technology in the Future

Jake Leach (Vice President, R&D, Dexcom, San Diego, CA)

Mr. Jake Leach updated the audience on Dexcom’s R&D efforts, starting with “the state of the art” – the G4 Platinum. He spent the bulk of his presentation covering the G4 Platinum’s strong accuracy data, which we have covered in detail at many recent conferences (links below). Mr. Leach also ran through the company’s impressive product pipeline, though he shared no new timelines beyond what we’ve heard in previous Dexcom earnings calls: a pediatric indication for the G4 Platinum (“we are eagerly awaiting approval”); Tandem t:slim and Animas Vibe pump integration partnerships (the Vibe is under FDA review; the t:slim is slated for completion “soon”); Dexcom Share (we got a first look at the Dexcom Share System and how users will be able to invite up to five people to remotely view their glucose via an App on their iPhone or iTouch); the new G4 Platinum algorithm (a remote web update is expected); the SweetSpot data management system (audience applause for Mac compatibility); the Gen 5 mobile platform (in addition to the smartphone, sending data to a watch is also under active development); and Gen 6 (potential for reduced or no calibrations). Mr. Leach concluded his presentation by noting Dexcom’s R&D goal: making products that are easy, accurate, and reliable (“The best sensor is the sensor that patients wear” – we agree).

  • To read our coverage of the G4 Platinum’s accuracy, please see page 73 of our ADA 2013 Full Report (, pages 128-130 of our EASD 2012 report (, and pages 29-31 of our ADA 2012 report (
    • The G4 Platinum’s improvements in accuracy are partially due to its redesigned sensor membrane. Mr. Leach reviewed the three key things sensor membranes must be designed to address: 1) limit the influx of glucose (you don’t want oxygen to become a limiting factor); 2) a biocompatible sensor interface; and 3) reduce the effect of interferences (either exogenous like acetaminophen, or endogenous like uric acid).
    • In developing the G4 Platinum, a big focus was on reducing the frustrating “???” errors sometimes seen with the Seven Plus. In these cases, the Seven Plus couldn’t figure out what the glucose was because of the underlying sensor signal. The device would display no CGM data (“???”), frustrating many patients. Dexcom discovered this was due to the microenvironment around the sensor resulting from inflammation and the foreign wound response. The new membrane reduces the occurrence and impact of these issues at insertion.
  • Mr. Leach reviewed Dexcom’s strong pipeline of products – as a reminder, Dexcom expects to release five new products in the next 12-18 months. Management emphasized this new strategy on the company’s 1Q13 call, which calls for a steady stream of incremental improvements rather than longer stretches between product improvements. We think this is brilliant, especially given the regulatory environment. 
    • G4 Platinum pediatric indication: Dexcom filed this with the FDA in February, and approval is expected before the end of the year. Current discussions concern the label’s accuracy tables, since the pediatric pivotal studies did not use the same glucose range in all age groups. Notably, this indication will expand the G4 Platinum to patients from 2-17 years old. The study tested alternate sites – both the upper buttock and abdomen – and the use of additional adhesives [SkinTac and Mastisol]).
    • Dexcom G4 Platinum pump partnerships (Animas and Tandem): The Animas Vibe is currently under review by the FDA, and the t:slim is a “very active development” that Dexcom is “planning to complete soon.” As of Dexcom 1Q13, an Animas Vibe launch in the US was estimated in 4Q13 or early 2014 (at Animas’ discretion), while FDA filing of the Tandem was expected before the end of 2013.
    • Dexcom Share: Before launching into a discussion on Dexcom Share, Mr. Leach asked the audience, “How many of you are familiar with Medtronic’s mySentry? Well, Share is very different.” The biggest change is that Share will be truly “remote.” Mr. Leach ran the audience through the basics: The G4 Platinum receiver will plug into a Dexcom Share docking cradle. The Share cradle (plugged into a power outlet) will both charge the receiver and transmit CGM data every five minutes to a nearby iPhone or iTouch via Bluetooth.  The Dexcom Share app on the nearby smartphone will receive the CGM data and send it to the cloud. Once there, the data can be shared with up to five people. Patients will be able to invite individuals by sending an invitation from the Share app on their iPhone. Those invited will receive a special link to download the Dexcom Follow app from the App Store. These followers will then be able to receive notifications of glucose excursions and view glucose trends on demand. The notifications will be completely programmable – e.g., a follower could choose not to be notified of a 200-mg/dl excursion after breakfast, but to get all hypoglycemia notifications. Mr. Leach did not share any timeline updates; as of the last update, a PMA supplement filing was expected in 3Q13. We have no doubt parents will love this device, especially if they have school-age children with diabetes. Management has also described Share as a key intermediate step towards the Gen 5 mobile platform, both from Dexcom’s perspective (developing a backend system, getting regulatory experience) and FDA’s perspective (getting comfortable with this area).
    • New G4 Platinum algorithm: Mr. Leach discussed the newest news from Dexcom’s 1Q13 report – a new algorithm for the G4 Platinum that will improve the overall MARD up to two percentage points (from 13.2% to 11.2%). The algorithm’s main accuracy improvements come from improved calibration algorithms that help reduce the number of outlier sensors. Mr. Leach showed an example of one sensor that began over-reading blood glucose following a calibration. With the new algorithm better interpreting the data, the outlier accuracy of this one sensor (MARD: 29%) improved to 10%. Mr. Leach did not share any timeline updates; as of the last news in Dexcom 1Q13 call, an FDA filing was expected in late 2013/early 2014.
    • SweetSpot data management system: Mr. Leach briefly detailed the goal of the web-based SweetSpot system: to integrate FDA cleared blood glucose meters, FDA cleared pumps, and Dexcom CGM into a single reports. SweetSpot will be Mac compatible, a fact that received a round of applause from the audience. There were no concrete timeline updates on this product; at Dexcom’s product theater at AACE 2013 in early May, Dr. Claudia Graham stated in Q&A that Dexcom is “submitting this year for a cloud-enabled system that can download to an EMR.”
    • Gen 5: We appreciate Mr. Leach’s commentary on Gen 5, which has the ambitious goal of sending CGM straight to the smartphone (it will use the same G4 Platinum sensor). Once data is on the phone, it can be seamlessly sent to the cloud. A key step in this R&D work is developing a communication interface that works with multiple manufacturers of cell phones – we had thought this was ANT+, though wonder if Dexcom will switch to Bluetooth. We were excited to hear Mr. Leach mention that Gen 5 doesn’t just have to talk to a phone; it could also send data to a watch. Encouragingly, that product is in active development. An ability to view CGM on a watch strikes us as very smart from a consumer perspective – we expect some users might have phones that will be incompatible with Gen 5’s communication technology, while others might prefer the more easily accessible wristwatch to view their CGM data.  
    • Gen 6: This closely followed R&D program is developing a brand new, “very highly accurate” sensor with improved interferent blocking (e.g., acetaminophen). “It’s an active program,” said Mr. Leach. It has potential for “reduced calibrations” and “maybe no calibration at all.” Dexcom will present data at future conferences. The latest timeline update we heard estimated launch of this product in 2017.


Data Overload!? Simplified Strategies for Retrospective CGM Interpretation and Assessment

Keri Weindel, MS, RD, CDE (Director, Professional Education, Dexcom, San Diego, CA)

Ms. Keri Weindel discussed Dexcom Studio, the new download software for the Dexcom G4 Platinum, as well as being backwards compatible, launched in October 2012. Ms. Weindel’s presentation was a blend of two previously presented at Dexcom’s product theater at AACE 2013 (see pages 55-58 of our report at for more detailed coverage). She began her talk by highlighting the two main design objectives of Studio: to “make the complicated simple” and to bring clinical value to practices – lofty goals indeed! Ms. Weindel’s focus was on the Dexcom Portrait report, a one-pager that really highlights the key problem areas of focus in bright red (hypoglycemia) and bright yellow (hyperglycemia). The software automatically identifies patterns to help make some sense of the classic CGM “spaghetti chart” that often overwhelms clinicians. The algorithms prioritize trouble areas by clinical significance based on the frequency, duration and intensity of hypoglycemia and hyperglycemia. Studio also provides a list of suggestions to help providers and patients devise strategies to address these patterns. We were encouraged by Ms. Weindel’s discussion of reimbursement, which noted that 98% of individuals with type 1 diabetes who have commercial insurance have personal CGM coverage, and coverage for type 2 diabetes is growing dramatically (albeit from a low base). It’s nice to see that Dexcom has specifically designed Studio to help HCPs fully document reimbursement for CGM interpretation (95251) – we’ve heard again and again that payment for CGM data interpretation is a challenge. In concluding, Ms. Weindel still cautioned that this software should not “supercede the interaction and interview with the patient.”

  • Dexcom Studio identifies clinically significant hypo- and hyperglycemic events using an algorithm that balances the frequency, duration, and intensity of the event. The greater amount of time spent outside of target range, the distance from in-range, and how often this event occurs influences how much clinical significance the algorithm places on the event. This also helps HCPs avoid focusing on eye-drawing outliers. We think Dexcom has made some great progress on this front.
  • To help explore hyper- and hypoglycemic events more closely, the microview part of Studio (Glucose Trends report) allows HCPs to see when patients had their meals and when they took their insulin – this, of course, depends on patients manually inputting these events into their G4 Platinum. Given the challenges of manual logging and the number of button pushes this takes on the receiver, our sense is that very few patients do this in practice.


Panel Discussion

Terry Gregg (CEO, Dexcom, San Diego, CA); Jake Leach (Vice President, R&D, Dexcom, San Diego, CA); Jay Skyler, MD (University of Miami Miller School of Medicine, Miami, FL); and Keri Weindel, MS, RD, CDE (Dexcom, San Diego, CA)

Mr. Gregg: We’ve got a lot on our plate. From a corporate standpoint, over the next 18-24 months, we have five new products that will be introduced. Our pipeline is extremely full. We focus on doing one thing exquisitely, and that is sensor technology.

Dr. Marian Rewers (Barbara Davis Center, Aurora, CO): I would like to congratulate the three speakers for putting together a really terrific program. It’s the best program I’ve seen so far at this meeting. One specific question about Dexcom Studio. This premise is appealing. But as a provider, sometimes there are some unusual cases you would like to discard. Can you manually remove days easily?

Ms. Weindel: Yes, you can eliminate certain days when they are an outlier – for example, you could select only weekdays. That algorithm was something we spent lots of time on.

Q: You can set that software to download or display 7 days 14 days, or 30 days. If you are trying to extrapolate, you take 30 days of data. But if you are trying to dive into details, you would go with seven days or day to day.

Ms. Weindel: I think when we look at clinical interpretation, sometimes it’s all about the now, or the most current information. As a clinician, we want you to take what’s meaningful for you. You can choose only weekends or weekdays and view as many days as you would like – there is lots of customizability in the software.

Dr. Skyler: The first thing I look at is the success report. Have they done better these past three months? Then I look at the 30 days, then the seven days. Are they taking the actions that we taught them? You can actually use all the parts of it separately and at the same visit. It depends on how well educated and experienced the patient is.

Q: I think that we should aim to train all our patients to read their own graphs, especially patients with new cases of diabetes. Are there any resources we can use to teach them?

Dr. Skyler: I train my patients to read their own graphs.

Ms. Weindel: We want to help educate patients in simple way. We have designed a training for the Dexcom G4 Platinum that is FDA approved for self-training. This is the first product that has been approved by the FDA with a tutorial to help patients use the device safely. Additionally, we are providing ongoing education. We have designed Beginning Your Journey, which is available online. This helps teach patients what the first week with CGM is like. We also have an education guide for patients who are looking to discover advanced features after they have used the CGM for, say, a month. This publication is called Discover Your Journey, and it is also available online. We aim to provide education that is meaningful to patients and physicians so that it reduces the burden for education on the device and its features and leave the CGM interpretation and use of real-time data to the clinicians. We want to give information that allows patients to dive deeper into the microview of self-management.

Mr. Leach: One thing I would like to add is that we are looking at education as we move forward to the mobile system. The Share app really explains how the system works so you really understand it as a user. It teaches you how to set up the system, how to invite people to support you, and many of the other features. This is an important part of our future.

Mr. Gregg: We are working with IDC to come up with better tools that are user friendly from both the physician and the patient standpoint. [Editor’s Note: We are very excited to hear that Dexcom is working with IDC on this front; we presume this is the Ambulatory Glucose Profile, and will be interested to see if the new SweetSpot system includes this one-page output.) You are all experts, but the bulk of people caring for patients with diabetes are not experts. I applaud the IDC efforts to streamline information in a more open architecture. Because of this, you can download one relatively simple report that is fairly uniform – I think in the future, industry will behave in more cohesive manner.

Q: I noticed the graphs had lots of reds and greens. What about color vision-challenged people?

Mr. Leach: That’s a constant discussion – what colors to pick for hypoglycemia. We originally had hyperglycemia as red, but we completely changed it so red is hypoglycemia. We don’t plan on changing the colors; we want them to remain consistent for all users

Q: Three questions: 1) Have you considered smaller sized data points on your printout? 20 When do you anticipate pediatric approval? 3) How will current G4 Platinum users upgrade to the new algorithm?

Mr. Leach: On the smaller dots on the printout, yes, that’s something we’re thinking about. On the new algorithm, the intent is that it will operate on the current receiver, and it will be a web upload onto those devices when it is approved.

Mr. Gregg: Regarding the pediatric indication, we filed that in 1Q13. Traditionally those PMA supplements are a 180-day review. We are in the last stages of it. All data has been analyzed and all the sites have been audited. Andy Balo has had multiple discussions with the FDA. I haven’t had an update today, but there was a discussion today. It’s all right now about labeling. In the original G4 Platinum trial that Jake referenced, we did 9,000 matched pair points. We drove those patients high and low so we could get the expansion into the number of hypoglycemia points: 15% of the 9,000 points are 70 mg/dl or below. That’s what FDA wanted. We cannot do that with two-year olds, so we didn’t have the same criteria. What came out of that was 157 patients in the pediatric trial. We are in discussions with the FDA on the tables. We cannot say the accuracy is X in 2-17 years old. We have to segregate those. That’s currently what we’re going back and forth on. My guess is sometime before the end of the year.

Q: In the pediatric studies, was the sensor worn for seven days?

Mr. Leach: Yes, they were seven days. The only change we made was the duality of application to the abdomen and to the upper buttocks – it was only an expansion on the points of administration.

Q: Could you elaborate on the interaction of the sensors with acetaminophen?

Mr. Leach: The Dexcom G4 Platinum is contraindicated for use with acetaminophen. With the Dexcom Seven Plus, the glucose signals read very high when patients took acetaminophen. The Dexcom G4 system has much less response to acetaminophen, but it is still contraindicated because of the rigorous testing the FDA has recently put in place. Our future systems are being designed to achieve interference blocking. They are going to have much less interference than the Dexcom G4 Platinum.

Dr. Skyler: As a user, I can say that when I was on Dexcom Seven Plus, I had to turn my sensor off when I took acetaminophen. With the Dexcom G4 sensor, I can take a Tylenol and have no problem at all.

Mr. Leach: That is not unusual at all from what we have heard.

Mr. Gregg: The Dexcom Gen 6 is still in very early stages. If you look at our critical care intravenous sensor in collaboration with Edwards, it is sensitive to 32 drugs. We are applying these lessons to Dexcom Gen 6, and we hope it is going to be highly resistant to a large drug panel. Our original goal was to develop the sensor for other areas in the hospital outside of critical care; but why can’t you do that in volume and make that the sensor of choice? In the future, our sensors will be more accurate and patients will hopefully not have to do the confirmatory fingerstick and recalibration, potentially with Dexcom Gen 6.  

Dr. Richard Bergenstal (International Diabetes Center at Park Nicollet, Minneapolis, MN): Thank you for an excellent program, and an amazing focus on more accurate sensors. I have a question on research. I’m pleased that almost every new drug is being studied with CGM. Where in the pipeline, is it Gen 5 or Gen 6, where if someone just wants to test and compare new drugs, they can use blinded CGM and it will go straight to the cloud?

Mr. Gregg; It’s here. We deal with probably 25 different pharmaceutical companies throughout the globe. We blind the sensors for them. Right now we’re in 18 different countries doing those clinical trials. Our corporate sales group seeks those contracts. That was part of the rationale for acquiring SweetSpot. The other company we’re associated with is Qualcomm. They have a device called the gateway. It has a bunch of radio chips in it. In the pharma world, the patient can connect the gateway, and it transmits the CGM data from the receiver to a cloud-based server. We perform the data analytics and then we port it to the sponsor. We do that today, Rich.

Dr. Robert Ratner (Chief Medical and Scientific Officer, ADA, Alexandria, VA): Very few people understand the difference between efficacy (an ideal situation) and effectiveness (real world) better than Jay. Tonight you said that you talked to Medtronic and asked them why they don’t link the Dexcom CGM to their pump. For closed loop, you want the best sensing system you can possibly find. But given what Dr. Hans DeVries talked about earlier today on liability, why is Dexcom involved at all in the closed loop?

Dr. Skyler: That’s a direction that the diabetes world is moving in. We’re making great progress, and I’m particularly impressed with the Boston group. They really are taking it out into the real world and finding that it works. That will really change the lives of people with diabetes. That’s control that will minimize and totally prevent complications.

Dr. Ratner: What happens with the first car accident when the pump doesn’t adjust to the CGM reading?

Dr. Skyler: I was one of the first people doing fingerstick monitoring in the days of urine glucose testing. The first time I presented data, it was in pregnant women. I was at Florida ADA affiliate meeting, the very organization you are the Chief Medical Officer of. I was a young guy – this was back in 1976 or early 1977.  I presented the data to a panel of three experts that were to critique it. The first, Karl Sussman, said that patients would not do it. They have to stick their finger; they’re not going to do it. We did not have automatic lancets at that time; it was using an insulin needle. Phil Felig said patients could not do it – they had to calibrate the device, it took 30 minutes to warm up after being plugged in, they had to time the blood sample with a stop watch, put the strip in, etc. They’re not technicians and they could not do it. Finally, Leonard Madison said patients should not do it. They’ll be out on the interstate and crash their car. I’ve got to tell you Bob: we don’t use urine testing anymore. They will do it, they can do it, and they should do it. All those guys were wrong. And there haven’t been many car accidents. And the accidents that happen are not due to equipment malfunction. Rather, it’s patients not paying attention. I find that in patients with hypoglycemia, they either take insulin and skip their meal, or they have excess exercise and don’t eat enough to compensate. It isn’t because the devices don’t work. I’m perfectly willing to have Dexcom in all artificial pancreas studies. It’s not device liability that’s the issue. If Ed Damiano’s calculations are right, we’ll have A1c’s of 6.1% with the bionic pancreas. That means no complications, normal lifespans, and normal quality of life. That I think is worthwhile. [Applause]

Q: I have a question regarding future advances with downloadable software. SweetSpot is trying to download all pumps and meters. Do you have an idea when we will get pump download data on your report?

Mr. Leach: SweetSpot has finger stick meters approved, and we are working with different pumps. We hope these advances will come in the near future – we are working on it.

Mr. Gregg: With SweetSpot, we are rapidly going to file it as a Class III medical device – that’s a PMA. There is a big difference between Class II and Class III devices. Anything that touches the CGM is a Class III device, and it is part of a PMA. The rigor that must go into a Class III PMA is quite extensive and is much, much greater than the work going into Class II devices. We have had to educate those at Animas and Tandem about the world they are getting themselves into. There are a lot of criticisms of the 510(k) process. But there is not a lot of comment about the PMA system – that’s because we are held to such a high standard. If you bring a new pump to market with a 510(k), it costs sponsors $75-100 million to get it to market. For a CGM with Class III PMA, that cost goes up to $250-300 million just because of the additional clinical studies and rigor. Here at Dexcom, we spend $15 million a year just on quality systems.

Q: Do you have plans to downsize the transmitter?

Mr. Leach: With every generation, the transmitters will get smaller – that’s with the Gen 5 program, and we’re looking into making the G4 smaller too. That’s an important part of patients’ convenience, and we want it to be as small as possible.

Mr. Gregg: We’ve set forth the criteria of a bottle cap. We’ve had to wait for that technology to evolve. Bluetooth LE has delayed a lot of what we do. All of you have Bluetooth. You cannot get through a single day on a phone. Bluetooth eliminates that. As we have this convergence of healthcare information, it takes enormous power. We sometimes have to lag behind the handset manufacturers, who lead all the development.

Q: Will that be a beer bottle cap or a milk carton bottle cap?

Mr. Gregg: Let’s go with the beer bottle cap. But don’t ask me – I’m not the engineer.


Industry Sponsored Dinner (Sponsored by BD)

How Sharp is your Knowledge?

Larry Hirsch, MD (BD Diabetes Care, Franklin Lakes, NJ)

Dr. Larry Hirsch opened the evening session with his history with diabetes, including his experience with glass syringes and boiling cases for sterilization. He shared with the audience his scholarship to college and medical school due to his grades and his “severe disability.” Dr. Hirsch made it clear, however, that he has lived his life without ever letting diabetes hold him back. He then turned to injection techniques, making a case for use of smaller needles – they provide the same level of glycemic control, less pain, and less risk of intramuscular injection. Quoting BD’s extensive research on the topic, he noted that skin thickness only varies from 1.25 -3.5 mm, allowing 4 mm needles to deliver insulin into subcutaneous tissue in all patients (for more on this, see page 17 of our report at Dr. Hirsch also discussed several news updates and areas of focus for BD at ADA 2013: use of 4 mm needles in obese individuals (posters 979-P and 997-P), the new EasyFlow needle with thin wall technology, and lipohypertrophy. To read more from our visit to BD’s ADA exhibit hall booth, see page 417 of our report at Dr. Hirsch closed his talk with an important line: injection technique matters.

  • A study presented at ADA compared different needle length in obese patients with diabetes. The study was conducted in adults with BMI >30 kg/m2, and was a crossover between 4 mm and 8 mm needles and 4 mm and 12.7 mm needles. Results indicated slight improvements in A1c levels with 4 mm needle use. Participants using the 4 mm needle reported no pain or leakage.
  • Dr. Hirsch also presented results of the new thin wall technology, which increased insulin flow, reduced dispense time, and reduced thumb force in bench studies (p <0.05) – this was our favorite demo in the ADA 2013 exhibit hall. Additionally, patients showed greater confidence, overall preference, and less time and thumb force to inject across Eli Lilly, Novo Nordisk, and Sanofi pens (p <0.001 for both pooled and individual results). The needle is now on the market.
  • Dr. Hirsch also briefly discussed lipohypertrophy, emphasizing that 94% of individuals who do not rotate injection sites have lipohypertrophy (vs. 5% in patients who rotate sites). He also mentioned a Spanish economic analysis that demonstrated a €122 million annual loss to the healthcare system due to lipohypertrophy. This raises the question of what the cost might be to other countries- we can imagine the US spends far more given the higher number of patients with diabetes.

Questions and Answers

Q: This whole idea is based on the premise that insulin delivery in subcutaneous adipose tissue has a different profile than that in intramuscular tissue. What is the evidence for that?

A: Dr. Hirsch is going to go over that, but without going into detail, there is very, very good evidence that absorption in intramuscular tissue is markedly changed if the patient remains at rest vs. if they exercise. It alters the blood flow to the tissue – increased exercise increases blood flow and increases absorption. This introduces variability that you do not want to have to deal with.

Q: In people with congenital lipohypertrophy, what kind of needle do you use?

A: That is a different animal than injection-related lipohypertrophy. I have yet to see anyone who has congenital lipohypertrophy. I will defer to my more learned colleague. The better answer is they need the appropriate medical therapy so they can try to build the subcutaneous adipose tissue.


Diabetes Injection Cases

Irl Hirsch, MD (University of Washington School of Medicine, Seattle, WA)

The second Dr. Hirsch of the evening used a series of cases to illustrate current hot topics in insulin administration. He started with lipohypertrophy [fat accumulation at the site of insulin injection], which affects a quarter to a half of type 1 diabetes patients in the modern insulin era but (in Dr. Hirsch’s opinion) has not been studied sufficiently. He mentioned anecdotal reports that insulin malabsorption due to lipohypertrophy has resulted in erratic glycemic control and severe hypoglycemia in some patients. He next discussed the effect of insulin dose on absorption, citing evidence that insulin is less effective in more obese type 2 diabetes patients (although whether the phenomenon is due to increased insulin resistance or increased adiposity is currently unknown). Dr. Hirsch struck down the theories that more frequent changes in infusion sites, abdominal versus thigh injections, or intramuscular injections can increase the response times of fast-acting insulin analogs – the one theory that evidence has borne out is the use of many small insulin injections over a single large injection (which is unfortunate from a patient convenience perspective, but very important that patients understand this phenomenon). Finally, Dr. Hirsch noted that shorter pen needles are not associated with increased leakage, as some providers have feared.

Questions and Answers

Q: Can you talk about the angle of the infusion sets in pump patients? In particular, 30 degrees versus 45 degrees?

A: I think it depends a lot on the length of the infusion set. It’s something I know of no data on. The bottom line is that we do what is most comfortable for the patient, because if the patient is uncomfortable they will get frustrated and won’t use it.

Q: We always think that insulin acts like a growth factor, which is what leads to lipohypertrophy. My question is whether we see the same thing with growth hormone.

A: I’m not sure; we would need to ask a pediatrician.

Comment: We had a patient in our practice on growth hormone that needed to have surgery on her abdomen because of the injections.

Q: Is there an increased risk of lipohypertrophy with increased doses in type 1 patients, or with infusion sets being used longer? Which is more likely?

A: We didn’t see data about changing the infusion sets, but that’s where I see the most lipohypertrophy. We get responses from questionnaires that some patients change their infusion sets once a week. The big problem on infusion sets is that we don’t have enough data.


Diabetes Drugs

Keynote Speaker

The Menu of Current Evidence-Based Therapies for Prevention of CV Outcomes in Type 2 Diabetes

Hertzel Gerstein, MD (McMaster University, Hamilton, Ontario, Canada)

Dr. Hertzel Gerstein kicked off day two of the conference with a well-organized, evidence-based evaluation of therapies for the prevention of cardiovascular outcomes in type 2 diabetes patients. Landmark investigations of intensive glucose lowering indicate that it yields modest, non-significant CVD benefits in advanced type 2 diabetes (e.g., ACCORD, ADVANCE, VADT), and can reduce long-term CVD risk and mortality in new-onset type 2 diabetes patients (e.g., the long-term follow up to UKPDS). In contrast to an earlier talk given by Dr. Robert Ratner, Dr. Gerstein did not voice misgivings about the large number of CVOTs currently being conducted, and instead seemed excited about the wealth of concrete data they will provide. Studies comparing the effects of different glucose-lowering strategies demonstrate that insulin sensitization and insulin provision approaches have similar effects on CVD risk. Dr. Gerstein noted that current evidence does not suggest that SFUs are particularly harmful (at least in a cardiovascular sense). He also touched upon the outcomes of recent trials demonstrating CV neutrality of some antihyperglycemic agents, such as ORIGIN and SAVOR. Regarding CANVAS, he did not dwell on the interim data presented earlier this year (which indicated an increased risk for MACE+ in the first 30 days, but was largely seen as a fluke), and instead looked ahead to more interpretable future data. In conclusion, he emphasized the importance of an evidence-based approach in dealing with the often-murky issue of cardiovascular risk in diabetes. Notably, during Q&A, Dr. Gerstein emphatically expressed the need for a randomized CVOT for gastric bypass surgery.

  • The incidence of cardiovascular disease (CVD) in diabetes patients has gone down, but a marked difference remains between diabetes patients and nondiabetic individuals. Meta-analysis of existing data shows a two to three times higher risk for CVD in diabetes patients, with a hazard ratio of approximately 1.5 per 1% rise in A1c. Dr. Gerstein argued that the source of these disparities must be associated with glucometabolic abnormalities, and that it is important to investigate whether glucose-lowering drugs can mitigate CV outcomes.
  • The intensity of glucose-lowering therapy has only a modest effect on CVD. Meta-analyses of four of the most respected intensive control trials in type 2 diabetes (UKPDS, ACCORD, ADVANCE, and VADT) show only a 9% reduction in CV risk. Long-term follow-up of the UKPDS showed a more distinct improvement.
  • Dr. Gerstein provided a rapid-fire survey of CVOT trials on different glucose-lowering agents. He noted that the long-term UKPDS follow-up showed a 27% reduced risk in CV-related death with metformin treatment. He expressed dismay at the lack of CVOTs focused on SFUs, but argued that current data does not indicate any significant CV safety risks with the drug class. He mentioned the neutral findings of ORIGIN and SAVOR (which investigated insulin glargine and saxagliptin, respectively) and also mentioned that AleCardio had been stopped (AleCardio was one of two CVOTs for the PPAR-alpha/gamma dual agonist aleglitazar, which Roche just chose to discontinue on July 10, 2013). However, Dr. Gerstein did not discuss results in depth – presumably, AleCardio was neutral, since Roche had been aiming for CV superiority. Regarding the CANVAS CVOT on the SGLT-2 inhibitor Invokana (canagliflozin), he briefly mentioned the interim data that was released to the FDA canagliflozin advisory board (read our full coverage here However, he did not dwell on those findings, noting the problems posed by trying to interpret interim data. In general, he felt that we don’t yet know enough about CVD risk with SGLT-2 inhibitors, incretins, or PPAR agonists.
  • Some therapies affecting cardiovascular comorbidities have been shown to reduce CVD risk. The BARI 2D study showed that coronary artery bypass grafting (CABG) reduced CVD risk, but not mortality, compared to intensive medical therapy alone in patients with type 2 diabetes and heart disease. Dr. Gerstein argued that the control of blood pressure down to 135 or 140 mmHg has accepted benefits; evidence does not conclusively support targets any lower than this. He also noted that fibrates added to statins do not reduce CVD and mortality, although he acknowledged that they have value in treating diabetes complications such as retinopathy.
  • Current RCT data on lifestyle intervention does not show a marked reduction in CVD risk. Dr. Gerstein cited the findings of the Look AHEAD trial, which demonstrated significant weight loss and other improvements in the treatment arm, but no accompanying reduction in CVD risk. He argued that providers and patients should adjust their expectations about lifestyle intervention, but not ignore the multiple other benefits a healthier lifestyle can confer.

Questions and Answers

Q: Is there any evidence about gastric bypass and cardiovascular risk?

A: That study must be done. Absolutely must be done. We need to do a study where we randomize patients and try them on gastric bypass, any form of gastric bypass. There is evidence from the Swedish Obesity trial, but it is NOT a randomized control trial. It needs to be tested in a large outcomes trial, and it’s something that the NIH should be interested in.

Q: What about fibrates? Where do you stand?

A: I think it’s clear that if you give a good dose of statins, fibrates aren’t needed. But fibrates are interesting because they show a clear benefit on diabetic eyes complications. Perhaps that presages a prevention on complication further down the road, but we don’t know about that yet.


Lessons Learned From the ORIGIN Trial

Hertzel Gerstein, MD (McMaster University, Hamilton, Ontario, Canada)

Dr. Hertzel Gerstein’s presentation of the ORIGIN trial results at ADA 2012 was one of the most widely anticipated talks in the last decade. We might now call him Dr. ORIGIN, as he has presented on lessons learned from the trial at multiple recent conferences since ADA 2012. The controversy over insulin’s use in type 2 diabetes is nearly as old as insulin therapy itself, he noted, and opinions on the issue have swung back and forth even in the recent past. Dr. Gerstein made it clear that ORIGIN was not intended to compare more vs. less intense therapy for diabetes; rather, it was primarily an investigation of the impact of insulin (glargine) use on cardiovascular harm. Moving through the results shared in 2012, Dr. Gerstein noted that glargine had no impact on CV outcomes or any type of cancer. However, data released this year from an ORIGIN sub-group (GRACE; n=1,091) shows that the glargine group saw modest, marginally significant decreases in Maximum Common CIMT and Bifurcation CIMT. Dr. Gerstein emphasized that the weight gain seen was modest, and that the risk of severe hypoglycemia, although significantly increased, was still only 1 event per 100 patient-years in the insulin glargine group. He was enthusiastic about the finding that insulin glargine treatment reduced the progression to type 2 diabetes by 28%. Dr. Gerstein noted that the data support the “beta cell rest” theory, although he made it clear that he does not advocate placing all patients with IFG or IGT on insulin. In conclusion, he stated that insulin glargine should continue being prescribed as needed, without undue fear of CVD or cancer. For our complete ORIGIN report from the 2012 ADA, see


Plenary: Gut and Diabetes and CVD Outcome Studies

Update on GLP-1 RA and DPP-4 Inhibitors

Robert Ratner, MD (American Diabetes Association, Alexandria, VA)

Dr. Ratner took the stage to provide an update on incretins, focusing largely on recent research and safety concerns regarding their non-glycemic effects (especially the issue of incretins, pancreatitis, and pancreatic cancer). He cited recently presented early evidence that incretins may have anti-inflammatory and neuroprotective potential. He emphasized the large number of CVOTs currently being conducted on incretins (which will result in an estimated 350,000 patient-years of data). Regarding the issue of incretin therapy and thyroid tumors, Dr. Ratner provided the familiar argument that these fears were drawn from studies on rodents, which have a high density of C-cell GLP-1 receptors as compared to human C-cells, which have no GLP-1 receptors. Dr. Ratner relayed much of the information presented at the NIDDK/NCI workshop on incretins, pancreatitis, and pancreatic cancer. This included a wide range of relevant preclinical, observational, histological, and randomized control trial data. He stated that although there are potential mechanisms that could connect incretins and pancreatitis, there is not enough concrete clinical evidence to justify an alteration in the use of incretin therapies. He did note that further study on the issue is needed, and that it is important to determine the most appropriate animal model for mechanistic investigations.

  • The research community is beginning to focus more on incretins’ non-glycemic direct effects on targets throughout the body. Dr. Ratner mentioned a recent wealth of studies investigating the anti-inflammatory actions of both GLP-1 agonists and DPP-4 inhibitors. Early evidence also suggests that incretins might be neuroprotective. One phase 2 clinical trial that is active but not yet recruiting is investigating the use of exendin-4 to treat Parkinson’s Disease ( Identifier: NCT01174810). Another phase 2 study, currently in the recruiting phase, is investigating exendin-4 in Alzheimer’s Disease patients ( Identifier: NCT01255163). A third study on the impact of liraglutide treatment on Alzheimer’s Disease ( Identifier: NCT01469351) was completed in April, although no data has been released as of yet.
  • Dr. Ratner briefly touched upon the issue of cardiovascular safety, noting that there are a large number of incretin CVOTs in progress. If all the studies are conducted as planned, they will yield a remarkable 350,000 patient-years of data. Dr. Ratner questioned whether the scientific and regulatory communities are asking and answering the right questions with CVOTs when it comes to incretins and cardiovascular risk, a question we certainly feel is valid since trial design (choice of endpoint, length, comparator, etc.) and enrollment (risk-enriched or new-onset) seem to have such a large bearing on the interpretation of results.
  • Dr. Ratner provided a familiar assessment of thyroid tumor fears. He noted that the studies that gave birth to the C-cell concerns were performed on rodents, which have a high density of GLP-1 receptors as compared to human C-cells, which have no GLP-1 receptors. He acknowledged a 2011 study that produced a slight C-cell pathology signal in the liraglutide treatment arm, but stated that this was most likely noise. His conclusion was that there is no convincing evidence of increased calcitonin release from C-cells or other C-cell abnormalities with incretin treatment.
  • Most of the talk was dedicated to a thorough, methodical analysis of pancreatitis and pancreatic carcinoma concerns. Dr. Ratner was an attendee at the NIDDK summit on incretins and pancreatitis (read our coverage of day one at, and our coverage of day two at He noted that this segment of his presentation was drawn extensively from Dr. Vanita Aroda’s session at ADA, with her permission (read our coverage of her presentation on page 121 of our ADA full report: He laid the groundwork for the discussion by noting that diabetes itself is known to increase patients’ risk for pancreatitis, before moving onto a survey of relevant preclinical studies. Although studies in healthy rodent models have shown no connection between high levels liraglutide (60-times higher exposure than doses used to treat diabetes) and pancreatic abnormalities, a mechanistic evaluation using the pancreatitis-prone Pdx-1 Kras mouse model showed that GLP-1 expanded pancreatic duct glands and increased mucinous metaplasia. Dr. Ratner also mentioned three post-marketing studies on pancreatic toxicology in a mouse models with elevated blood glucose and triglycerides, none of which demonstrated treatment-related adverse effects on pancreatic exocrine histology or proliferation. He underscored the importance of finding the optimal animal model for further mechanistic investigations, as healthy animals are a poor simulator of diabetes patients.
  • Evidence from RCTs and histological studies have also failed to produce convincing evidence connecting incretins and pancreatitis. An important point of consideration is that rare events such as pancreatic carcinoma are unlikely to be picked up by RCTs, which are generally limited in size. Dr. Ratner spent a great deal of time discussing Dr. Peter Butler’s morphological study on human pancreases, which initially seemed to indicate a connection between incretin therapy and pancreatic cell proliferation. However, Dr. Ratner went on to point out a number of limitations of the study, including the poor matching of the two groups by age and diabetes duration, possible inclusion of type 1 diabetes patients, and the use of a polyclonal antibody that was later shown to stain nonspecifically.
  • The ADA, Endocrine Society, EASD, and IDF have all recommended that the current data are not enough to recommend a change in prescription behavior for incretins. The ADA, EASD, and IDF released a joint statement to this effect earlier this month – read our take on the document at Dr. Ratner ended by framing the pancreatitis issue with four succinct questions: 1) Are there potential mechanisms to explain the safety issues (yes); 2) Is there concrete clinical evidence to suggest an alteration of current risk/benefit profiles of these agents (no); 3) Are there any substantial changes in clinical recommendations based on the studies (no); and 4) Is there a need to pool together appropriate expertise, databases, and methodologies to address these questions with scientific rigor (an emphatic yes!).


Basic Physiology Pathogenesis of Diabetes

Robert Rizza, MD (Mayo Clinic, Rochester, MN)

Dr. Robert Rizza opened his compelling presentation underscoring that understanding the pathophysiology of type 2 diabetes is critical for developing better antidiabetic agents and improving our use of the ones already available. Dr. Rizza believes that with a deeper understanding of type 2 diabetes, physicians could determine ways to use agents synergistically rather than just complimentarily. Dr. Rizza proceeded to clearly describe the disease’s pathophysiology by breaking it up into seven question and answer segments, which covered a broad spread of topics (details below). We were particularly interested in his explanation of the differences between IFG and IGT and the implications these variances might have – people with IGT have abnormal insulin secretion and action, yet people with IFG are normal for both and instead have excessive hepatic glucose production. Thus, Dr. Rizza suggested that IFG and IGT might be different diseases and therefore require different treatment strategies (he did not describe what these could be). Throughout these different Q&A sections, Dr. Rizza emphasized that excessive hepatic glucose production is responsible for elevated fasting glucose levels in both type 2 diabetes and prediabetes. He thereby implied that targeting hepatic glucose production could be a particularly promising treatment strategy. In concluding his presentation, Dr. Rizza proposed that an ideal treatment for type 2 diabetes would 1) normalize glucose production and glucose uptake before and after food ingestion, 2) normalize fasting and early postprandial insulin secretion in a glucose dependent manner, 3) normalize insulin action in all tissues, 4) normalize glucose effectiveness, and 5) normalize glucagon secretion.     

  • While displaying his disclosure slide, Dr. Rizza championed the importance of academic researchers working with pharmaceutical companies. He asserted that he has learned a “tremendous amount” working with industry in part because companies have models largely unavailable in academia. He, therefore, called his conflicts of interest, “duality of interests.”
  • Dr. Rizza broke his explanation of type 2 diabetes pathophysiology into seven question and answer sections.  

1. What is the cause of fasting and postprandial hyperglycemia in people with type 2 diabetes? Dr. Rizza detailed how both fasting and postprandial hyperglycemia are caused by elevated hepatic glucose production in combination with the lack of a compensatory increase in glucose uptake (Firth et al., JCI 1986).

2. Are people with type 2 diabetes insulin resistant? If so, what tissues are insulin resistant? Yes. Impacted tissues include but are not limited to the liver (excessive endogenous glucose production and reduced glucose uptake), visceral organs (reduced glucose uptake), and muscles (reduced glucose uptake) (Basu et al., Diabetes 2004). Dr. Rizza stated that at physiologic insulin levels, reduced glucose uptake in the muscles account for roughly two-thirds of the type 2 diabetes body’s decreased glucose uptake. The liver’s drop in glucose uptake accounts for the other third. Dr. Rizza noted that peripheral tissues and liver tissue have very different glucose uptake profiles, meaning addressing insulin resistance in different tissues could require different treatment strategies.

3. Do people with diabetes and prediabetes have alterations in insulin secretion? According to Dr. Rizza it depends on the type of prediabetes. People with diabetes or IFG/IGT have abnormal insulin secretion and action; however, people with IFG have normal insulin secretion and action. People with IFG have elevated glucose levels due to excessive endogenous glucose production (Bock et al., Diabetes 2006). Given the differing pathophysiology of IFG and IGT, Dr. Rizza suggested that they might require different therapeutic strategies to treat. 

4. What is the cause of fasting hyperglycemia in people with diabetes? Excessive endogenous glucose production. Notably, however, Dr. Rizza’s group found that in the basal state endogenous glucose production is only elevated among people with “severe” diabetes (basal glucose level of ~12.7 mmol/l) – people with “mild” diabetes (basal glucose level of ~8.1 mmol/l) have a comparable production rate to people without diabetes. Yet, in the clamp state, people with “mild” diabetes  (~11.4 mmol/l) also have a significantly elevated rate of endogenous glucose production (Basu et al., Am J Physiol 2004). Dr. Rizza hypothesized that this trend could result from hepatic glucose production not being adequately suppressed by high glucose levels.      

5. What is the cause of fasting hyperglycemia in people with prediabetes? Excessive endogenous glucose production. In a study Dr. Rizza’s group conducted, people with IFG/NGT had higher fasting glucose production due to increased rates of gluconeogenesis. People with IFG/NGT’s gluconeogenesis was suppressed but still trending high during a hyperinsulinemic clamp. Additionally, Dr. Rizza stated that people with IFG/NGT and IFG/IGT had impaired insulin suppression of glycogenolysis (Basu et al., J Clin Endocrinol Metab 2013).    

6.Does a delay in insulin secretion have a different effect on postprandial glucose tolerance than does a decrease in insulin action? Yes, a delay in insulin secretion results in higher peak postprandial glucose concentrations. Whereas, a decrease in insulin action results in prolongation of postprandial hyperglycemia.

7.Is “glucose effectiveness” (i.e., glucose’s ability to regulate its own metabolism) impaired in people with type 2 diabetes? Yes; however, Dr. Rizza has not seen any data on drugs’ abilities to improve glucose effectiveness. Dr. Rizza’s group caused healthy people and people with type 2 diabetes to have flat insulin levels by holding their basal insulin levels constant with an insulin infusion and suppressing endogenous insulin secretion with somatostatin. Researchers then infused participants with 35 g of glucose. People with type 2 diabetes’ resulting blood glucose levels were almost 50 mg/dl higher than that of people without diabetes who underwent the same procedure. Thus, Dr. Rizza concluded that glucose’s ability to regulate its own metabolism in the presence of basal insulin concentrations is abnormal in people with type 2 diabetes (Basu et al., Diabetes 1997). 


CVD Outcome Studies in Type 2 Diabetes – How Did We Get Here and Where Are We Going?

Angelyn Bethel, MD (University of Oxford Diabetes Trials Unit, Oxford, UK)

Dr. Angelyn Bethel provided a basic overview of cardiovascular outcomes trials for type 2 diabetes. She emphasized from the start that for cardiovascular disease, blood pressure, lipids, antiplatelet agents, and smoking cessation are the most important factors to control (“there is a small role for glucose”).  Her presentation reviewed data from landmark trials, highlighted the 2008 FDA cardiovascular guidelines, ran through current CVOT studies for several drug classes, and concluded with a discussion of off-target effects (e.g., pancreatitis). She believes the future of outcomes trials in diabetes will include more complex trials, which will require efficiencies in trial design and conduct (an increasingly discussed option is the “pragmatic clinical trial,” a more real-world RCT). Dr. Bethel believes that quantifying off-target effects of diabetes drugs is the next challenge; while she believes “common” outcomes like cardiovascular disease are “easy to study,” rare outcomes will require new and innovative approaches.

  • Dr. Bethel discussed how the 2008 FDA guidelines have greatly impacted trial design in diabetes drug research (Bethel et al., Curr Cardiol Rep 2012). She searched the database for diabetes drug CVOTs, comparing the March 2005-March 2008 period to the March 2008-March 2011 period (the FDA cardiovascular guidance came out in March 2008). “The number of CVOTs are growing in number, size and complexity,” she noted – indeed, there were twice as many CVOT trials pre- vs. post-guidance (eight vs. 16), median trial size increased by a factor of six (n=1,116 vs. 6,000), the number of trial sites increased “exponentially” (160 vs. 501), and global distribution broadened to more countries (17 vs. 27; particularly Latin America and Africa). In her view, the one good thing that came out of this guidance was alignment of cardiovascular composites: nearly all studies now include cardiovascular mortality, non-fatal MI/ACS, non-fatal stroke, and hospitalization.
  • “If you want to bring a new drug to market, unless you have an overwhelming cardiovascular effect size early in your clinical program, you’re going to be doing a large outcomes trial.” Dr. Bethel first touched on the ongoing CVOTs for GLP-1 agonists: ELIXA (lixisenatide), EXSCEL (exenatide), LEADER (liraglutide), and REWIND (dulaglutide). Broadly speaking, she noted that most are enrolling similar patient populations, and the variable part is what level of cardiovascular risk is being included. She displayed DPP-4 inhibitor trials as well: EXAMINE (alogliptin), TECOS (sitagliptin), SAVOR TIMI-53 (saxagliptin), and CAROLINA (linagliptin). A following slide showed SGLT-2 CVOTs: BI10773 and CANVAS (canagliflozin). She concluded with the glitazones and glitazars, noting two ongoing studies: TOSCA IT (pioglitazone) and ALECARDIO (aleglitazar) – we note that the latter was an error, as Roche halted the ALECARDIO trial on July 10 (press release: We were disappointed to see that this was not acknowledged in her talk or noted on the slide.
  • Off-target drug effects are generally unexpected and rare events; while the gold standard is the randomized controlled trial, this method of study is not always feasible. The main problem with RCTs is the rarity of off-target events, making statistical significance difficult to achieve – this is even the case when thousands of patients are studied over many years. Dr. Bethel mentioned a few other methodologies (epidemiology, post-marketing surveillance, and registries/EMRs), though these approaches have their own inherent problems: incomplete data, an unknown total number of exposed patients, reporting bias (new drugs, media attention), and little information to assess exposure time and relatedness.


Panel Discussion

Moderator: Robert Ratner, MD (American Diabetes Association, Alexandria, VA)

Panelists: Angelyn Bethel, MD (University of Oxford Diabetes Trials Unit, Oxford, UK); David Dunger, MD (University of Cambridge, Cambridge, UK); Robert Rizza, MD (Mayo Clinic, Rochester, MN)

Dr. Ratner: Are large-scale trials really appropriate to answer these safety [pancreatitis] questions?

Dr. Bethel: In order to answer this question I have to compromise my own job security. I think the requirement of large-scale approaches is not going to be sustainable in the long run from a financial standpoint, and also the science doesn’t support it. We’ll have to be much more careful when looking at adaptive trial designs or other ways to look at these questions.

Dr. Jay Skyler: Are these trials going to be adjudicated enough such that we can use them for answering the pancreatic question?

Dr. Bethel: They were not designed for that. Given the low rate of pancreatitis and pancreatic cancer I do not think we will be able to get a definitive answer from a single trial. With the ridiculous number of patient years we will have though, we might be able to get an answer from pooling the results. The question is how much we trust such an analysis; the trials vary a lot in how they adjudicate these events. For example, we are running two trials. In one, the pancreatic event is described and adjudicated – period. The other has a more complex process for event review. As we heard earlier, there are biases in the conduction of these reviews, which are difficult to overcome.

Dr. Rizza: First I want to disclose a personal conflict of interest: I am very close friend with Dr. Peter Butler, as are our families. The issue with cancer is whether this is an accelerator or not, and I don’t know the answer. We need a properly controlled and properly powered study on pancreases. Even if what Angie is saying is done, you’ll have a tremendous number of patients exposed, and if anything this is an accelerator of a very rare event.

Dr. Bethel: You need to remember that none of these trials are collecting pancreas tissue trials.

Dr. Rizza: Right but you should be able to find enough pancreatic tissue samples to do such an analysis. This can be done properly. Frankly, I don't know what to do right now.

Dr. Bethel: You can follow the advice that has been given: provide the best care you can to your patients and explain the data so that they can understand. This is difficult to do, since the data is difficult to wrap your mind around. Still you have to work to ensure that your patient is a partner in the decision.

Dr. Rizza: Absolutely.

Q: For Dr. Rizza, two quick questions: is prediabetes actually just an earlier form of type 2 diabetes? And regarding the pathophysiology, is the gluconeogenesis response a survival trait?

Robert Rizza: I think that when diabetes was actually changed to “type 2” from “adult onset,” the idea was that type 2 diabetes is a syndrome that could have a number of causes. I think the same thing is true for prediabetes. I think that we’ll find that the cause of IFG is a variety of things, but different than the causes of IGT. Regarding gluconeogenesis, I think that if you have a drug that specifically targets glucose uptake, it might be the wrong drug. And what’s happening in transient diabetes, say in youth, is that there are children for which SFUs can cure the disease. We don’t know that right now. It’s important to understand the pathogenesis so that when more drugs become available, we’ll know the best way to use them.

Dr. Ratner: Dr. Dunger, we know that statins work well and are safe in adults. We also know that they can be very dangerous during pregnancy. What do you do with teenagers, a group in which there is no such thing as a planned pregnancy?

Dr. Dunger: That is anxiety provoking. How do I say this? The patients who are getting pregnant at our clinic are not the ones we would enroll in such a trial.


Plenary: Management of Adult Diabetes

Is Hypoglycemia an Issue in Type 2 Diabetes?

Stephanie Amiel, MD (King’s College London, UK)

Dr. Stephanie Amiel’s focused presentation directly addressed the title of her talk. She built a strong case that hypoglycemia is important in type 2 diabetes (whether that’s mild or severe), though she made it clear that there is a lower risk in those with early type 2 diabetes and residual insulin secretion. Most notable was Dr. Amiel’s discussion of the recent ADA hypoglycemia guidelines (Diabetes Care 2013) – she was uncomfortable with the statement’s use of <70 mg/dl as a hypoglycemia threshold AND an “alert value.” In her view, using 70 mg/dl as the definition of hypoglycemia is likely to “grossly overestimate” the amount of clinically important hypoglycemia. She also discussed data from ADVANCE and suggested that hypoglycemia is a marker for morbidity and mortality, not necessarily a cause of either (Zoungas et al., NEJM 2010). Last, Dr. Amiel listed several factors that predict who is at risk for hypoglycemia: history of severe hypoglycemia, longer diabetes duration, loss of endogenous insulin secretion, co-morbidities, awareness status, and use of intensive therapy/lower A1c (the latter is controversial). She believes that the risk for clinically important hypoglycemia should be assessed individually.

  • Dr. Amiel had a mixed take on the new ADA guidelines on hypoglycemia (Seaquist et al., Diabetes Care 2013). Specifically, she took issue with the use of <70 mg/dl as both a definition of “hypoglycemia” and an “alert value.” While Dr. Amiel was a fan of the alert value approach, she felt that <70 mg/dl as a definition of hypoglycemia was likely to overestimate the amount of hypoglycemia experienced in studies (“Some of us are very uncomfortable with this”). This contrasts with the EMA’s former 2006 position, which defined hypoglycemia as <3 mmol/l (<54 mg/dl). Dr. Amiel lamented that the revised EMA 2012 guidelines recommend using the ADA consensus statement. In Q&A, Dr. Hans DeVries mentioned that Dr. Amiel has pushed the EMA to lower the threshold.
    • For context, the new guidelines document characterizes hypoglycemia using five different terms: 1) severe (requiring assistance); 2) documented symptomatic (measured low blood glucose); 3) asymptomatic (no symptoms but a measure low blood glucose); 4) probable symptomatic (typical symptoms but no measured low); and 5) pseudo hypoglycemia (typical hypoglycemia symptoms but the glucose is only approaching low).
  • Hypoglycemia IS an issue for people with type 2 diabetes, especially those on insulin. Dr. Amiel reviewed a few studies, all showing clinically meaningful rates of severe hypoglycemia in people with type 2 diabetes – in the case of type 2s on insulin, rates even mirrored those with type 1 diabetes. Dr. Amiel’s review of the literature suggested very broad ranges for severe events in type 1 and type 2 diabetes: 3.8-170 episodes per 100 person years for type 1 and 0.4-44 episodes per 100 patient years for type 2. A 2007 publication in Diabetologia from the UK Hypoglycemia Study Group prospectively studied hypoglycemia in type 1s and type 2s, finding equivalent rates of ~30% of type 2s on insulin (>5 years of therapy) and type 1s (<5 years of diabetes duration) reporting at least one severe hypoglycemia. Rates were around 10% for type 2s on sulfonylureas and type 2s on insulin (<2 years of therapy). Nearly 50% of those with type 1 for >15 years reported at least one severe hypoglycemia event.
  • Mild hypoglycemia is important! Dr. Amiel covered a survey (Brod et al., Value in Health 2011) that asked patients to describe their experiences after non-severe nocturnal hypoglycemia. A striking 23% of patients reported missing work time after such episodes. Interestingly, this varied by country – Germans and Americans reported much lower rates of missing work, while patients from the UK and Germany “took any excuse to miss work.”

Questions and Answers

Q: What is pseudo hypoglycemia?

A: I have only read the paper. It’s what they used to call relative hypoglycemia. It’s the perception of hypoglycemia when you are absolutely not hypoglycemic. For people with chronically poor control, they will experience symptoms of hypoglycemia at normal glucose levels.

Comment: There was an important message from ACCORD about cognitive function and severe hypoglycemia. The message was that cognitive dysfunction predicted severe hypoglycemia – it wasn’t the reverse. Another point from ACCORD is that severe hypoglycemia does not cause severe outcomes. It’s a marker for people at risk for them.


Intensive Diabetes Management in Type 2 Diabetes

Robert Rizza, MD (Mayo Clinic, Rochester, MN)

In his presentation on management of type 2 diabetes, Dr. Robert Rizza discussed the importance of understanding how drug mechanisms complement each other in order to help an individual achieve glycemic control – we have noticed more and more KOLs starting to tout the importance of combination therapies for this reason (and many, like Dr. Rizza, have been talking about combination therapy for a very long time). Dr. Rizza remarked that while we currently have a glucocentric approach to diabetes therapy, it may be beneficial to move to a holistic approach (e.g. expanding to also emphasize lipid control and amino acid turn over) because, Dr. Rizza noted, normalizing glucose does not guarantee normalization of these other factors that are dysregulated in diabetes. Dr. Rizza began his discussion of diabetes treatment with lifestyle modification. Significantly, he commented that the termination of the Look AHEAD trial does not mean that lifestyle modification is ineffective, but, rather, the Look AHEAD results demonstrate that patients need to dedicate to sustained, significant lifestyle modifications (for more information, please see page 20 of our ADA 2013 report at In his discussion of sulfonylureas, Dr. Rizza touched on the issue of mortality, remarking that he would “like to point out the only agent that results in higher mortality is insulin.” Turning to GLP-1 agonists, Dr. Rizza highlighted that there have been no long-term outcome studies showing reduced microvascular or macrovascular events, although he has his “fingers crossed” that the results will be positive. In response to the question of whether or not it matters that GLP-1s suppress glucagon, Dr. Rizza says yes and no – he stated that it only becomes significant if it is paired with a drug that increases insulin secretion (presumably because the combination could produce hypoglycemia), corroborating his message of considering how drugs complement each other in combination. Dr. Rizza also noted that DPP-4 inhibitors have positive effects on inflammation and immune function, though they are not as potent as other drugs. Remarking on SGLT inhibitors, Dr. Rizza commented that it is important to consider that SGLT-1 inhibitors may have off-target effects on other tissues. 

  • Dr. Rizza emphasized individualized treatment. He noted that there can be much inter-patient variability in response to certain drugs (for example, there is a large inter-patient variability with glargine use, and there is even large intra-patient variability with insulin use), and we can expect true improvements in diabetes care when drugs whose benefits outweigh the risks in specific individuals are used in a complementary fashion.
  • “Not all sulfonylureas are the same” Dr. Rizza highlighted, describing one study that showed that glimepiride, but not glyburide, affects pre-ischemic conditions in type 2 diabetes (Tung-Ming, JCEM 2003). In contrast, he also noted that sulfonylurea use has been shown to be associated with reduced mortality over the long-term in a follow-up of the United Kingdom Prospective Diabetes Study (UKPDS) – however, the UKPDS was not designed as a study of sulfonylurea efficacy, so in our view, we are not sure it is completely fair to draw this conclusion.


How Big an Issue is Hypoglycemia in Type 1 Diabetes?

Hans DeVries, MD (University of Amsterdam, Amsterdam, Netherlands)

Dr. Hans DeVries began his presentation by explaining that even though the majority of patients with type 1 diabetes (about 70%) do not experience an episode of severe hypoglycemia (SH) in a year, it is a large problem in the patients it does affect. He emphasized the duration-dependent nature of SH, remarking that hypoglycemia unawareness develops when patients no longer produce as much adrenaline or glucagon in response to hypoglycemia. Turning to outcomes after SH, Dr. DeVries noted that patients experiencing a SH event have a higher mortality rate. He noted that while it is hard to attribute the mortality solely to SH, one study has shown that, after controlling for other factors, a patient who had experienced a SH event had a 3.4-fold higher mortality rate at five years (McCoy, Diabetes Care 2012). On a positive note, Dr. DeVries called data from DCCT “reassuring,” noting that there was little change in cognitive function between patients who experienced no SH events and those that have experienced more than five (although Dr. DeVries did note that those with A1c levels of >8.8% did experience a decline in two cognitive domains). Turning to how technologies may help improve patients’ hypoglycemia, Dr. DeVries noted his own, online analysis comparing the effects of subcutaneous insulin injections and multiple daily injections that was in response to Fatourechi’s meta-analysis performed on the same topic. He remarked that his analysis demonstrated that pump treatment can diminish the rates of hypoglycemia. Additionally, Dr. DeVries noted that although patients using low glucose suspend (LGS) technology did show less time in hypoglycemia, and the overall accuracies of CGMs are improving (the Roche prototype has a MARD of 9%!), he would still caution against viewing CGMs as a comprehensive answer to hypoglycemia: there was no difference in SH events in patients using LGS compared to the control group.  He expressed his concern that one study using Guardian REAL-Time CGMs, the MARD dropped from 17% to 39% when patients fell into hypoglycemia, and the positive predictive value for the hypoglycemia alarm was 42% (Zijlstra et al., Diabetes Care Obesity and Metabolism 2013). Finally, Dr. DeVries commended Dr. Trang Ly on his study comparing the reduction of SH in hypoglycemia unaware patients on either LGS or the pump alone. The data were very compelling, with SH events dropping to zero. To read more about Dr. Ly’s study, please see page 63 of our ADA 2013 report at


Panel Discussion

Hans DeVries, MD (Academic Medical Center, Amsterdam, Netherlands); Stephanie Amiel, MD (King’s College London, UK); Robert Rizza, MD (Mayo Clinic, Rochester, MN); Abhimanyu Garg, MD (UT South Western Med. Center, Dallas, TX)

Dr. Robert Ratner (ADA, Alexandria, VA): Since neither of you were on the EASD/ADA panel for hypoglycemia, where would you set the threshold for clinically relevant hypoglycemia?

Dr. Amiel: It is contextual. Hypoglycemia is anything under 3.5 mmol/l [63 mg/dl]. I am interested in preventing the sort of hypoglycemia that is associated with cognitive dysfunction. And the avoidance of which is related to restoration of awareness. That’s 54 mg/dl.

Dr. DeVries: Stephanie took initiative to convince he EMA to bring it down. I joined the initiative. I completely agree with her.

Dr. Garg: Is the threshold for hypoglycemia different in men vs. women? My understanding is that women can tolerate hypoglycemia much better without getting cognitive dysfunction.

Dr. DeVries: We men don’t tolerate anything better than women.

Dr. Rizza: it turns out that women can. If you fast a woman for 72 hours, the mean glucose is 38 mg/dl. The mean glucose for men is 45 mg/dl. Woman have larger brains for their body mass. The brain is glucose dependent. I do not think 70 mg/dl is the right number. I think it’s 55 mg/dl. I like what the ADA did – I think 70mg/dl is an alert value. If the glucose is less than 70 mg/dl, it is lower than normal.

Dr. Amiel: It’s semantics. The new document moves in the right way – 70 mg/dl is an alert value. But then a few paragraphs later, it says this is hypoglycemia. That is the way it’s being interpreted and used.

Dr. Rizza: 70 mg/dl is hypoglycemia. Normal people do not get below 70 mg/dl – not unless you give them a pure glucose load, or unless you fast them for a long period of time. I think this is more than semantics. Lower than 70 mg/dl is not a normal glucose concentration. If you get there consistently, that’s a problem.

Dr. Satish Garg (Barbara Davis Center, Aurora, CO): Many individuals come to the clinic and say, ‘I’m absolutely fine when my blood glucose is in the 30s or 40s.’ What is the answer to those individuals? I show them these red bars all the time on the CGM. They absolutely say, ‘I have no problem. I do not want it to go any higher.’

Dr. Amiel: If you put them in front of a test of cognitive function. They will not perform appropriately. These people are hypoglycemia unaware. When you actually look at what happens in the brain, these people do not get the appropriate changes in the cognitive area of the brain. They don’t know they are hypoglycemic. They don’t think there is anything wrong in being down there. We have found that for some of these people, you cannot convince them that there is a problem. We have just started looking at this – using psychological techniques to engage these people. We’ve only done a very tiny pilot.

Dr. Abhimanyu Garg: I see many type 2s. I say the ADA normal fasting glucose is 70-130 mg/dl. But they say, ‘By the time I get to 100 mg/dl, I fell hypoglycemic.’ The counter-regulation starts in chronically hyperglycemic people at 100 or 90 mg/dl. What’s the answer to these people?

Dr. DeVries: Go down gradually and the counter-regulation symptoms will probably disappear. The problem Satish mentioned is really a problem. There is a small group of patients that idealize hypoglycemia. ‘As long as I’m hypoglycemic, I’m not hyperglycemic. And if I’m not hyperglycemic, I’m not getting complications.’ It’s a very tough group to give guidance to, and you’re often not successful.

Dr. Rizza; It’s a very serious psychological problem – people become absolutely obsessed. This is very serious. High is bad, lower, lower, lower. That’s time to refer to psychiatric care.

Dr. Amiel: We are all able to reverse hypoglycemia unawareness in the research setting. At least three or four groups have done this. It doesn’t last – that’s the problem. Out of the research setting, you cannot sustain the changes that they make.

Mr. Claude Piche (Locemia, Montreal, Canada): On the relationship between severe hypoglycemia and adverse outcomes, what is the effect of time spent in severe hypoglycemia? Is it a marker or a cause? I’ve never seen data that says, ‘These patients were in the severe hypoglycemia state for 30 minutes, 45 minutes, 2 hours. How long until they were retreated with glucagon at home, by an ambulance, or with glucose?’ Are there any data that associate severity of outcomes with time spent in the severe hypoglycemic state?

Dr. Amiel: Does anyone know?

Dr. Satish Garg: In data from the recent ASPIRE trial, you saw four individuals with severe hypoglycemia in the LGS off group – these were the only individuals in the trial that experienced severe hypoglycemia. Their time spent hypoglycemic was significantly higher – close to two hours. That’s the only real data that was available. We don’t know how long it takes for seizures or a coma.

Dr. Rizza: Unless you are continuously measuring, you don’t know. Hepatic glycogen depletes in 30-40 minutes. There is glycogen in the brain, but not much. As you go through the sequence, and probably are depleting every substrate you have, this is going to be a big problem. It’s an important research question.

Q: You mentioned that in type 1 diabetes, you consistently divide doses of basal insulin. Does that help with hypoglycemia?

Dr. Amiel: Our preferred policy for intensive flexible treatment in type 1 diabetes is a divided basal. We use NPH and Levemir, which are not 24-hour insulins. It provides flexibility. We can change the dose of basal inulin the night you’ve gone for long jog. When people come to us with a hypoglycemia problem, we retrain them on insulin therapy. That’s part of it – smaller amounts of insulin in two blocks rather than all at once. I’m not sure if I’ve ever looked at that one element in the context of producing hypoglycemia. We know the training course reduces hypoglycemia and restores awareness.

Q: What about splitting the dose in type 2?

Dr. Amiel: We haven’t looked at that. The protocols of the 4T study called for bedtime insulin with NPH. When that no longer holds them, if they have good fasting glucose, we add a morning dose or add in prandial insulin. We’ve never specifically compared to twice daily vs. once daily.

Dr. Rizza: We would commonly use two basal rates. The more injections, the more stability you have. We commonly have people doing this. Pick your favorite basal, glargine or detemir, and it lets you change more and smooth out the concentrations. If you look at the kinetics, there is less variability. Does that decrease severe hypoglycemia? I do not know the answer.

Dr. Ratner: Do you see any justification for regulatory approval of an antidiabetic agent that has potential of causing hypoglycemia? Will we see another agent, like a sulfonylurea, that stimulates insulin secretion in an unregulated fashion? I’m talking about a new drug class aside from insulin.

Dr. Rizza: Everything is risk-benefit. There could be agents that have a potent ability to suppress glucagon. There may be a small increase in hypoglycemia in a non-insulin dependent manner. I think it’s unlikely, because everyone is going to glucose dependent.


Plenary: Frontiers in Type 1 Diabetes

New Basal and Prandial Insulins

Robert Ratner, MD (American Diabetes Association, Alexandria, VA)

Dr. Robert Ratner provided a comprehensive overview of new basal and prandial insulins, as well as new insulin delivery modalities. After describing his ideal basal insulin (non-peaking profile, low plasma insulin variability, very low rates of hypoglycemia, and no weight gain), he began discussing Novo Nordisk’s Tresiba (insulin degludec). He was positive on the insulin’s pharmacokinetic, pharmacodynamic, and safety profile. Next in the presentation was LY2605541, Lilly’s PEGylated lispro, which has also looked impressive in early testing and might have preferential hepatic uptake (a mechanism that could reduce insulin-related weight gain). Dr. Ratner mentioned that new concentrated insulin formulations being tested by Sanofi (U300 glargine) and Novo Nordisk (U200 degludec) seem to have remarkably predictable pharmacodynamic profiles and less hypoglycemia. Turning to rapid-acting insulins, he discussed Biodel’s EDTA-citrate zinc chelator candidate BIOD-250 and hyaluronidase-based approaches for hastening insulin’s actions, along with Novo Nordisk’s FIAsp (which is slated to enter phase 3 testing later this year). He ended with a discussion of oral, buccal, and inhaled insulin delivery modalities, areas that are being aggressively pursued by a number of companies. With so many exciting new insulins in development, we were glad to hear Dr. Ratner’s take on the field of candidates.

  • New basal insulin candidates bring us one step closer to the “ideal,” which has a flatter PK/PD profile and lower rate of hypoglycemia. Dr. Ratner began with Novo Nordisk’s insulin degludec, which has a demonstrated 25-hour half-life, stable concentration profile over 24 hours, and a 25% reduction in nocturnal hypoglycemia. He mentioned the FDA’s February complete response letter (CRL) requiring further CV safety data for insulin degludec (read our report on the FDA’s decision, and noted that US CVOTs are planned. He then moved on to LY2605541, Lilly’s PEGylated insulin lispro candidate. As with insulin degludec, LY2605541 proved at least as effective as previous insulins, with a lower risk of hypoglycemia. Interestingly, studies in a dog model have suggested that the candidate may have preferential hepatic uptake, which (if confirmed in human studies) may have the potential to reduce insulin-related weight gain. The theory is supported by early data showing slight weight loss with LY2605541 treatment, but Dr. Ratner was clear that the mechanism behind the weight loss has not yet been confirmed. There are five ongoing phase 3 (IMAGINE) trials in both type 1 and type 2 diabetes comparing LY2605541 to currently available basal insulins, and data from phase 2 trial data is slowly being released. Dr. Ratner then discussed Novo Nordisk’s insulin degludec U200 and Sanofi’s preparation U300 insulin glargine formulations — the latter significantly reduced the incidence of hypoglycemia (especially nocturnal hypoglycemia) compared to the U100 glargine formulation.
  • Hyalouronidase-linked insulin and Biodel’s citrate/EDTA zinc chelators (BIOD-238 & BIOD-250) both have an earlier peak and smaller tail than currently available bolus insulins. By Dr. Ratner’s understanding, BIOD-250 is moving towards phase 2 testing. Recombinant human hyaluronidase went through extensive phase 2 testing two years ago, but updates since then have been scant. Dr. Ratner noted that these faster acting insulins would be a boon for closed-loop systems.  He also mentioned Novo Nordisk’s FIAsp, which should soon enter phase 3 testing in both type 1 and type 2 diabetes patients.
  • Alternative insulin delivery options are being aggressively investigated by a number of companies. Many modalities that initially seemed promising (transdermal, nasal, sublingual) have fallen by the wayside. The buccal insulin Oral-lyn did not prove non-inferiority in a phase 3 trial (read our take on the data at and was not able to achieve an acceptable dosage in less than four sprays (the FDA’s requirement). However, it is currently approved in Ecuador and moving through phase 3 trials in India. Dr. Ratner next discussed oral insulins, which he labeled the “holy grail” of insulin therapy. Biocon, Oramed, Capsulin, and Novo Nordisk all have oral insulin candidates in some stage of development. MannKind is the last man standing (so to speak) in the field of inhaled insulins. Dr. Ratner shared data showing that the company’s Technosphere Insulin is absorbed and cleared faster than any other insulin he has seen, so quickly that it may require mid-meal re-dosing. Studies of the candidate are ongoing, and results are expected in August or September.


Islet Transplants: Is it a Reality?

Ronald Gill, PhD (University of Colorado, Denver, Denver, CO)

Dr. Ronald Gill opened by dedicating his presentation to the late Dr. George Eisenbarth. Dr. Gill then proceeded to argue that the notion islet cell transplants don't work is a myth. He explained that a volume of islets the size of one’s thumb is enough to make a person insulin independent, and, with enough grafts a person can remain so for life. Dr. Gill believes that the Edmonton protocol was the “victim of its own hype.” He explained that the expectation that people would be insulin independent for life under the first protocol was unjustified and that the initial protocol was “not a great method in retrospect.” For example, it utilized rapamycin, which reduces beta cells ability to regenerate. According to Dr. Gill, the protocol has evolved and improved. Among people who received a graft in 1999-2002, 27% were insulin independent three years later. Among people who received a graft in 2003-2006, 37% were insulin independent three years later. Finally, among those who received islets in 2007-2010, 44% were insulin independent three years later. Thus, Dr. Gill compared islet transplants to the insulin pump – it continues to improve from the original “model.” Additionally, Dr. Gill pressed that persistent islet function, even without insulin independence, provides both protection from severe hypoglycemia and improved A1c levels. Looking to the future, Dr. Gill noted that the “elephant in the living room” is that we do not have a robust way to induce immune tolerance in humans.

  • The Clinical Islet Transplantation Consortium is conducting a multi-center demonstration trial for islet transplant alone or an islet transplant following a kidney transplant. More information on the trial can be accessed at  


Intensive Insulin Management Issues in Type 1 Diabetes

Jay Skyler, MD (University of Miami Miller School of Medicine, Miami, FL)

Dr. Jay Skyler gave an enlightening presentation on the history of intensive insulin therapy. Starting in the days prior to insulin's discovery (The Allen Plan's starvation diets), he walked attendees through the discovery of insulin; early insulin regimens with regular and NPH (24-25 gauge needles, standardized meals, urine glucose testing); the transition in 1976 to split-mixed insulin schedules; the first article on home blood glucose monitoring and basal-bolus therapy (Danowski and Sunder, Diabetes Care 1978); the first lancets (“this looks like it was designed like something out of a medieval torture chamber”); the publication of Intensive Insulin Therapy in 1983 (Drs. Skyler, Schade, Santiago, and Rizza met at the Chicago O'Hare Hilton once a month, with each writing a chapter and critiquing one another’s’ chapters; the book became the methods section for the DCCT); the DCCT from 1986-1993; the emergence of flexible insulin therapy with insulin analogs in 1995 (Humalog) and 2000 (Lantus); and finally, the critical importance of adding CGM. Said Dr. Skyler: “Intensive insulin therapy – we've come a long way.” Wow is he right.


Panel Discussion

Ronald Gill, PhD (University of Colorado, Denver, Denver, CO); Aaron Michels, MD (Barbara Davis Center for Diabetes, Aurora, CO); Robert Ratner, MD (CSMO, American Diabetes Association, Alexandria, VA); Jay Skyler, MD (University of Miami Miller School of Medicine, Miami, FL); Sir George Alberti, MD (Imperial College, London, UK)

Comment: You were talking about the potential benefits of a very flat once a day basal insulin. It is possible people do better on twice daily basal insulin. For example, if you are active you might want to take less at night. I am very puzzled as to why we think a once daily is what we want.

Dr. Ratner: There are a couple caveats on who should and who should not get it. The current basal insulins that we utilize, glargine and detemir are clearly not full 24-hour coverage. If you are dealing with type 1 diabetes, I agree with Jay – you use twice-daily therapy. The newer therapies get their advantages from consistently lasting a day. I did the new degludec trials. About 20% of my type 1 diabetes patients could not tolerate it because their exercise was not daily or not consistent. So in type 1 diabetes, I agree that you might not want a completely flat 24-hour insulin. However, in type 1 diabetes it might still be helpful to have that flat basal insulin – to have this long tail of basal insulin to get you through an illness is important.

Dr. Larry Hirsh (BD, Franklin Lakes, NJ): A brief comment. In terms of trying to look for more rapid prandial insulins, I don't think that you mentioned that there is another route that is intradermal delivery with very small needles. We are working on that at BD. You get almost the same kinetics as what you saw with Halozyme. You made a comment about the variability of more targeted insulins. I think that many of us get that. Are you aware of any other companies that are trying to target insulin more to the liver? That would seem to me to be even more beneficial than speeding it up. 

Dr. Ratner: I am also a fan of medical history. One of my favorite people was Peter Warshaw, who developed diabetes. He went to Harvard and his physician was Elliot Joslin. Peter complained that Elliot was not rigorous enough. Peter took six to eight shots of insulin per day and completely discounted the intermediate and long duration insulin. He stated that the appropriate way to take insulin was via the hemorrhoids, since they directly connect to the portal system. From an intellectual perspective, I think we can say that makes sense. We do not see much data showing the advantage of portal delivery. The only data I have seen on this difference is with the old pump studies where you either had it going into the subcutaneous space or the portal system, and you only saw some differences in lipids. To date, I don't think that there is any clinical data showing real differences between portal and subcutaneous delivery.

Q: A couple of years ago, I was at one of the Diabetes Technology Society meetings, and there was an Ohio company that presented on hepatic-directed vesicles that could be added to insulin. They showed some curves with impressive reductions in postprandial peaks. I’ve never heard anything about it since then.

Dr. Ratner: There’s a reason you haven’t heard about it since then.

Dr. Skyler: I should point out that when you give peritoneal insulin, only half goes portally. The other half goes systemic. I think that the other thing that was seen with the implanted pump experience was substantially less hypoglycemia, and I think that that is because we did get at least half going to the liver. The full experiment as you want it, Bob, hasn’t been done, and we need a portal specific insulin to do it.

Dr. Ratner: I agree, and I don’t think that the real advantage would be the insulin only going to the liver. I think the advantage is that it’s not hitting the systemic circulation in high levels. We're making all our patients hyperinsulinemic by injecting insulin systematically, but it's the only way we can get adequate effects. 

Dr. Skyler: There’s someone with a good quote on this — I think that was you George. What was the quote?

Dr. George Alberti: I’m on a statin so my memory is gone. [Laughter] The quote was: wrong place, wrong time, wrong dose. 

Q: A few years ago we heard about smart, glucose-dependent insulin. Any updates?

Dr. Ratner: The concept of linking insulin release to ambient glucose is a relatively old. The original idea was embedding insulin in a matrix so that it could be released stoichiometrically when the glucose levels rose. It didn’t work back in 1980. A lot of companies are working on it now; the idea of an insulin sponge is now being developed. Others are looking at nanoparticle-linked insulin, where there is an exchange of insulin for ambient glucose. Those studies are very early, and the earlier ones failed miserably.

Q: The Hylenex preparation I know has been used for quite a few things. In recruiting for the trial I have worries about using if for an extended period of time – two years – especially since it breaks down hyaluronic acid.

Dr. Ratner: That is a real concern. I am going to plead ignorance and ask the other panel members where things stand. We actually participated in the original trials of the hyaluronic acid products linked with the Hylenex prep, but have never seen the complete data.

Dr. Skyler: We actually had a call last week with the principal authors trying to get the paper out. It turns out that the hyaluronic acid in the subcutaneous issue is rapidly reconstructed. So it should not matter because you keep replacing it all the time.  I think they need a co-formulation to really make it work.


Plenary: Diabetes Complications

Update on Microvascular and CVD in Diabetes

Jay Skyler, MD (University of Miami, Miami, FL)

Although it is widely accepted that glucose levels are closely linked to increased cardiovascular risk, we lack concrete evidence that improving glycemic control decreases patients’ risk for CVD. Dr. Skyler explained this “glucose paradox” by citing potential limitations in three major trials that demonstrated CVD neutrality with intensive insulin treatment: ACCORD, ADVANCE, and the VADT. He noted that the studies were either short in duration, had insufficient power, had well-controlled patients in the comparator arm, or included patients who had longstanding diabetes and existing cardiovascular issues. The fact that meta-analyses of these studies does show a significant (albeit modest) CVD improvement with intensive treatment, combined with the 20-year findings of DCCT/EDIC that also show a CVD benefit, indicates that such studies must be allowed to progress further for the CVD benefit to emerge. Turning to the issue of risk factors for CVD in diabetes, Dr. Skyler noted that although trials have not shown an association between more intensive BP control and reduced overall mortality, there was a definitive decrease in patients’ risk for stroke. He emphasized the importance of distinguishing between different types of CVD, and aiming for lower BP targets if a patient has a family history of stroke. Statins, he said, have clear cardiovascular benefits for diabetes patients, while fibrates generally did not. He concluded by noting that reducing CVD complications in type 2 diabetes involves a number of factors, including glycemic control, blood pressure, lipidemia, and lifestyle choices. 

  • The period between 1999 and 2010 saw improvements in CVD in US diabetes patients. The NEJM study released in 2013 showed that over half of type 2 diabetes patients in 2010 achieved A1cs less than 7%, blood pressures below 130/80 mmHg, and LDL below 100 mg/dl (although very few patients achieved all three). Both diabetes patients and nondiabetic individuals have seen a decrease in CVD, but a difference in prevalence still exists between the two groups.
  • The “glucose paradox,” or lack of evidence linking better glycemic control to reduced CVD risk, might be explained by limitations of ACCORD, ADVANCE, and the VADT. Dr. Skyler noted that VADT was a relatively small study (n = 1791), and all three studies enrolled patients with longstanding diabetes (8 to 12 years), many of who had an established history of CVD. He noted that EDIC took 20 years to demonstrate a CV benefit, indicating that ACCORD, ADVANCE, and the VADT may have had positive results if they were allowed to proceed further, thus narrowing the data confidence intervals. Meta-analysis data from the three studies indicating a modest yet significant CV benefit support the idea that the studies may have been close to achieving positive results. Other potential limitations of the three trials include the fact that the standard treatment control arms were relatively well controlled in terms of A1c and blood pressure.
  • Blood pressure (BP) control has been proven to reduce CVD in diabetes patients, but the exact target remains a point of some uncertainty. The ADA 2012 guidelines set 130/80 mmHg as a goal, but the 2013 guidelines raised the target to 140/80 mmHg. This decision was informed by the result of the ACCORD BP trial, which showed no evidence of CV improvement below a systolic blood pressure of 140 mmHg. Dr. Skyler did note that other studies have shown a continuing improvement in CVD below 140 mmHg. Although many trials showed no improvement in CVD risk with intensive BP control, a significant difference in patient risk for stroke was demonstrated. As a result, Dr. Skyler said, lower targets might be advisable in patients with a family history of stroke.
  • Statins have generally proven to reduce CVD risk in diabetes patients. Dr. Skyler cited data from two meta-analysis conducted by the Cholesterol Treatment Trialists’ Collaborators in 2008 and 2012 which supported statin use. For comparison, ACCORD results indicated that fenofibrate does not significantly reduce CV events.


Panel Discussion

Sir George Alberti, MD (Imperial College, London, UK), Satish Garg, MD (University of Colorado, Aurora, CO); Robert Ratner, MD (American Diabetes Association, Alexandria, VA); Jay Skyler, MD (University of Miami Miller School of Medicine)

Q: Why did the ADA change its blood pressure goals, when they know that, given that people are living so long with the disease, it makes sense to lower blood pressure? What was the justification?

Dr. Ratner: The ADA Professional Practice committee had been waiting since 2008 to change the guidelines because of JNC [Joint National Committee] 8, because we wanted to be consistent with whatever JNC 8 said. When we were preparing the guidelines this year we gave up and said: let’s look at the data ourselves, regardless of what the NHLBI is going to do. Looking at the data, there is clear-cut evidence that reducing blood pressure to 140 mmHg has a benefit. Two studies looked at reducing blood pressure to 130 mmHg, and they showed no effect. The rationale was that if it’s difficult to get to a blood pressure of 130 mmHg, then you add more side effects by adding more medications. I agree with Jay that if you can easily get to 130 mmHg, then that makes sense. But there is no rationale for pushing more and more blood pressure meds to get to 130 mmHg

Dr. Skyler: The current ADA guidelines for glycemia are personalized. If that’s the case, why don’t we personalize blood pressure, accounting for family history and other factors?

Dr. Ratner: I agree, and what you are talking about is matching population health with personalized care, and that’s really the direction we’re moving in. We’re always at a disadvantage when we describe diabetes patients as a body that is all uniform. They’re not. People with diabetes have individualized goals.

Dr. Alberti: That should be the point of guidelines as opposed to protocols.

Q: I was wondering if you could explain the rationale for not using niacin?

Dr. Garg: I did one study a few years back with what I thought were moderately high grams of niacin (4.5 grams) – no one uses that value now. We showed that in the placebo-controlled trial, many patients became hypoglycemic and hyperuricemic. One of my patients experienced a rise in uric levels from about 7 mg/dl to about 11 or 12 mg/dl. We know that niacin can deteriorate glycemic control. Other trials have shown that modest doses of niacin, such as 1.5 or 2 grams, does not cause a great deal of glycemic control deterioration — but we do still see elevations in A1c levels. Because of this, I have been a little skeptical of niacin. Anecdotally, I can tell you patients with poorly controlled glycemic levels would come to me with triglycerides of 300 and 200 mg/dl, and when I stopped their niacin, their glycemic control would get getter. My enthusiasm for niacin is low.

Dr. Skyler: My diabetes is niacin induced. I use niacin to control my triglyceride levels. It pushed my A1c to over 6, and I figured I knew how to take care of diabetes, and niacin was the only thing working for my triglycerides. My glucose levels may be higher but I still want to reduce my triglycerides. Remember, you have to think about individualized, personalized care.

Q: This is from the previous talk, but the question is related to HDL. Should we not worry about HDL, keeping in mind that type 1 diabetes patients usually have higher HDL?

Dr. Garg: As you saw, the current guidelines don’t mention HDL, or even if they do, they mention it as a secondary target of therapy. They also mention that you should raise your HDL by means such as reducing saturated fat diets and exercising. The studies that have been done with CETP inhibitors have been negative.

Comment (Dr. Irl Hirsch, University of Washington, Seattle, WA): I have a comment about the personalization of blood pressure targets. I do not think that this is well captured in current ADA guidelines. The guidelines you are giving are more appropriate for patients with type 2 diabetes. If you have a 25 year-old man with a blood pressure of 135/80 mmHg and documentation of a previous blood pressure of 110/115 mmHg, that 135 mmHg is bad sign. In this situation, people need to appreciate what the guidelines mean for a type 1 diabetes population.

Dr. Skyler: This is an advantage with electronic records. They have great flow sheets for records such as blood pressure. If you see a change in baseline, and rise in a patient with type 1 diabetes in particular, that is an indication for intervention.

Comment (Dr. Hirsch): But that is not well captured in the guidelines.

Dr. Skyler: It does need to be captured more.

Dr. Ratner: We are progressing more toward talking with more patient-centered-ness. You can see that with the ADA/EASD type 2 diabetes guidelines. We are evolving in all standards of care. In the future, there will be very specific recommendations differentiating type 1 diabetes and type 2 diabetes because they are different. We are moving in that direction, and we are moving fairly aggressively. Let me use your example. If you look at current standard of care, a patient with microalbuminuria is to be on therapy regardless of blood pressure. The nice thing about microalbuminuria is that it is reversible so that if that 25 year-old has gone from 110 mmHg to 135 mmHg, next year they might have microalbuminuria, and you can initiate therapy. We know that both blood pressure and microalbuminuria will respond. I don’t think there is data for prophylactic use.

Comment: Does everyone understand all of the association between ACE inhibitors and the data for retinopathy? The randomized trials are complicated and to get these into the guidelines is complicated.

Dr. Ratner: This is directed at population health and incorporated to meet individual needs.

Dr. Hirsch: In patients with type 2 diabetes, we have all of these different choices, and you are talking about doing more thoughtful trials. We have to make sure we are using the right medications. TZDs have hazard ratios of CV mortality - there is no difference between pioglitazone [Takeda’s Actos] and rosiglitazone [GSK’s Avandia]. The real issue to me is not just glucose control, but what pharmaceutical treatments do we use to give us better advice. Right now, we are just throwing everything at patients without thinking it through.

Dr. Skyler: There was a discussion that pointed out that if you choose therapies for type 2 diabetes you have to balance glucose lowering, impact on weight, and ease of use—that’s the personalized use. What to choose is a separate question. What I was addressing was that lowering glucose has benefits in terms of outcomes. I think that when we achieve 7% A1c levels in the control arm in ACCORD, it makes it hard to say that glucose lowering below that is substantial.

Q: Can you comment on the association between cognitive impairment and statins?

Dr. Garg: As you know, the FDA introduced a change in the label of statins and recently added cognitive impairment. The experience we are seeing is anecdotal, and FDA reports on cognitive impairment are case studies. I have seen patients taking statins. One young lady would forget what she was supposed to do when she went to the computer. When I stopped her statins, her memory became clear. Patients who were 50 to 60 years old would forget where they were driving. In some cases, it appeared to be the statins. As a general rule, when I see patients with statins, I used to ask them if they have muscle pains and cramps. I ask them if they are sleeping well, or if they are experiencing any memory loss or forgetfulness.

Dr. Skyler: Some use statins to treat Alzheimer’s and are claiming successes. It is only statins that cross the blood-brain barrier.

Dr. Ratner: The other aspect is the very specific wording. It is not dementia. It is a different mechanism compared to Alzheimer’s. It is real, speaking as a study of one. When I meet you out in the hallway and I don’t remember your name, it is because of the statins [laughter].

Q: How effective do you think changing from a Western diet to a plant-based diet will be at correcting someone’s lipid panel, and would you pay a price in terms of glycemic control?

Dr. Garg: As I showed you, I’m not sure we can say if a vegetarian diet would be effective. There are no such studies. Also different people define “vegetarian” differently. I am vegetarian and drink milk, but more and more people are going vegan. A lot of the oil sources in plants are high in saturated fats. The one advantage is that plants don’t make cholesterol, so your cholesterol content in your meal will plummet. But saturated fat is the primary influence on serum cholesterol, and dietary cholesterol is second.


Plenary: Management of Type 1 Diabetes in Children

Intensive Management of Type 1 Diabetes

Moshe Phillip, MD (Tel Aviv University, Petah Tikva, Israel)

Dr. Moshe Phillip opened his presentation by rhetorically asking, "What is the alternative [to intensive insulin therapy]?" Dr. Phillip championed the use of pump therapy, calling it the "new conventional therapy for type 1 diabetes." Strikingly, despite Dr. Phillip's fervent passion for pumps, he lamented that 40-50% of his patients refuse to use one. Similarly, Dr. Phillip explained that few people use CGM despite it being associated with a significant decrease in A1c (particularly among those with the highest baseline A1cs and who use the sensor most often [Pickup et al., BMJ 2011]) and with a reduced time spent in the hypoglycemic range (but not a reduction in the incidence of severe hypoglycemia [Phillip et al., Pediatric Diabetes 2012]). According to a survey by the T1D Exchange, about 3% of people younger than 26 years and ~13% of those 26 years or older are on CGM. In the German and Austrian registry, DPV database, only 4.8% of patients (2.3% of pediatric patients) use CGM. Dr. Phillip hypothesized that low utilization is not due to poor reimbursement, since many patients do not use CGM even when they are given the sensor. He pressed that providers have to try and figure out why and encourage industry to find solutions. Dr. Phillip believes implantable sensors might be the solution. As Dr. Phillip touched upon his use of low glucose suspend pumps in Israel, we heard the attendee behind us sigh, "I wish we had that," followed by a dejected response from her neighbor, "I know." As a reminder, Medtronic recently received an approvable letter from the FDA for the MiniMed 530G. The last update in Medtronic F4Q13, targeted approval by the end of 2013 (for more details, please see our Medtronic F4Q13 report at


Prevention of Cardio-Renal Complications

David Maahs, MD, PhD (Barbara Davis Center, Aurora, CO)

Dr. David Maahs began by arguing that the medical community should be as concerned with cardiovascular and renal complications of type 1 diabetes as it is with glycemic control. He noted that persistent microalbuminuria (MA) occurs in 20-40% of patients, is a predictor of later cardiovascular risk, and is preventable. The management of cardio-renal complications, in his mind, is a major area of unmet need that presents opportunities to significantly improve patient outcomes. Instead of recommending a series of potential solutions, Dr. Maahs posed a number of difficult questions, including how aggressively providers should pursue treatment of cardio-renal complications over glycemic control, at what age treatment should begin, whether providers should set targets less stringent for children than those set for adults, and what balance between pharmacotherapy and lifestyle intervention they should aim for. Throughout the presentation, Dr. Maahs noted the lack of data on many of these questions — we certainly agree with him that these are important issues that definitely merit further study.


Panel Discussion

Marian Rewers, MD, PhD (Barbara Davis Center, Aurora, CO); Moshé Phillip (Tel Aviv University, Tel Aviv, Israel); David Maahs, MD, PhD (Barbara Davis Center, Aurora, CO); Jennifer Raymond, MD (Barbara Davis Center, Aurora, CO); Georgeanna Klingensmith, MD (Barbara Davis Center, Aurora, CO); David Dunger, MD (University of Cambridge, Cambridge, UK)

Q: Do you really believe that type 1 diabetes patients are at increased risk for cardiovascular mortality. If there is increased risk, what should we do about it?

Dr. Maahs: looking at the data from the Scotland Registry, they clearly have a two to three times greater incidence of early death compared to the nondiabetics. That same risk was present for cardiovascular risk in particular, and cardiovascular mortality.

Dr. Dunger: No way. We have been discussing this study for maybe 10 years now. There are markers to cardiovascular disease during adolescence, A1c levels are worse during adolescence. They are definitely the high-risk group. So what do you do? You allow adult treatment practices to come down to the adolescence. You begin using statins and other drugs. You have no ideas what the safety is and how long to keep them on it. That was the basis of the AdDit study. Until then I don't think we have any rational reason to prescribe treatment.

Dr. Rewers: When are results coming?

Dr. Dunger: End of 2015

Dr. Klingensmith: I agree with David. We desperately need data on this. I also think that family history may help to inform us. I think we need to remember to re-take that family history in adolescence. Is there now a history of hypertension or early heart disease? If so, and if their A1c is 10 and we cant get it down, and the child is somewhat hypertensive and has lipids in the 150-160 range, maybe we should consider therapy.

Dr. Dunger: Where you do see high lipid levels and high blood pressure, it’s usually because of a strong family history, so it might not be related to diabetes or microalbuminuria — it might be a matter of family history. And families with history may be more willing to try to treat the condition, because family members will have had treatment as well. Maybe there will be a family member who has started statins, so they decide that with a high LDL of 150 they want to begin treatment. And we don’t necessarily need to hit a LDL of 100 — we can just get them most of the way there. Also education of the patient and family is important, letting them know that this is an important risk factor.

Dr. Phillip: I think that in most cases it is really easy. If you have tremendously high lipids, blood pressure, or microalbuminuria you begin using the medications. The question is what to do wit those in the middle. For those patients I will wait two years for the results.

Q: I was shocked that nobody mentioned hypoglycemia in this session. What are your recommendations about hypo?

Dr. Phillip: We are aiming to prevent hypoglycemia in every age group. I’m actually optimistic about the future with the new tech that will become available, even before closed-loop systems.

Q: We know that high glucose also raises cholesterol so what do you do with a 15 or 16 year old girl who has an A1c of 9.5% cholesterol of 210 mg. What are your priorities?

Dr. Maahs: We are always going to try and intensify glucose control. We do have data from the SEARCH study what when we looked at the association between the change in A1c and change in the lipids, even if you get the A1c down to goal, you are not going to get the LDL down to 100 mg. Now you can debate whether or not you should treat with a statin. I don't think you should think that dropping an A1c from 9% to 7% will drop the LDL enough to get to goal.  

Q: I heard some voices at meetings like this about hypoglycemia-induced seizures in children who are ten or twelve.

Dr. Klingensmith: I have heard those same discussions. Hypoglycemia is dangerous; it frightens parents and it can cause death. In a town the size of Denver, everyone is destroyed when someone dies of hypoglycemia, including the physician community. Children can be intellectually damaged from a severe event. I don’t think we can minimize the importance of hypoglycemia events.

Q: Regarding shared medical appointments, while you are doing the appointments with the teens are you meeting with the parents to help them let go?

Dr. Klingensmith: We started with just the teens and we have added a SMA with the parents at the same time.

Comment: I had the pleasure of being at a SEARCH meeting immediately before Keystone. At the SEARCH meeting we recognized that we are about half way through our data collection for our cohort study. I agree that 2015 is so long to wait for the AdDit results. We have a group that will number about 3,000 who were diagnosed in 2002 to 2008. By the time we see them a second time they will have had at least five years of diabetes duration. We are looking for early markers of kidney disease, atherosclerosis, and diabetic retinopathy. Fortunately, we are a little bit ahead of our data collection. We have decided that we are going to come out with some early papers for the first 1,500 patients. We will be coming out with a number of papers and presentations at ADA and hope to have that by early June. Just a short advertisement.

Dr. Rewers: SEARCH is a very important study. These studies will paint a much better picture of what to be doing.

Dr. Klingensmith: I would add that our data on the impact of exercise is not new. Data about exercise has been very, very helpful.

Dr. Rewers: I think that Moshe’s work with the low glucose suspend and overnight control is so important. I think that all of these are going to reduce the amount of time below 60 and 70 mg/dl and it will be a tremendous benefit to patients and their families.

Dr. Klingensmith: I guarantee you that that the engineer was saying he wanted the MDLAP in his home not because he wanted to avoid hyperglycemia but because he wanted to avoid hypoglycemia.

Dr. Phillip: With the current technology available, we’ve reached what we can reach, and it’s not the target we want yet. We need better technology, and industry is working on it.

Q: About the shared medical appointments, do you have kids come together with one CDE for about an hour but before that they come in and you download the meters? Do you discuss the blood glucose levels with everybody or individually. How long are kids there?

Dr. Klingensmith: The blood glucose are analyzed by the provider sitting in the back of the room to answer any medical questions that come up. It is my understanding that they answer those questions individually.

Dr. Phillip: What if you adopt the age I use for the transition, which is 30?

Dr. Klingensmith: I think we need to understand the additional needs of the adolescents and the young adult and help them become real adults as quickly as possible rather than staying as emerging adults.

Dr. Rewers: I would not let you go without answering the question: what do we do in patients with persistent A1cs above 9%?

Dr. Klingensmith: I would love to hear from the audience what you do with those patients.

Dr. Dunger: In the UK, the group aged 18-25 are the most vulnerable. There is a high risk of particularly young men being found dead – not just dead in bed but just dead. You are leaving home and all of those contacts. We tried to do joint clinics with the adult endos but they are so busy the adult endos never came. We are doing a transition period through the age of 19 years. With this, some of the nurses will see the patients in the pediatric clinic and then later in the adult clinic.  So the nurses see patients in both clinics. It is an extensive solution.

Dr. Klingensmith: It is very, very impossible. A few things will work intermittently – a text from mom or dad but then mom and dad have to remember and that is difficult. And money is very effective which we can’t do but there families can.

Q: Some people are starting to study transitions, one of the studies out of Boston studied patients after their transition or lack thereof, and its very stressful for them. ENDO, ADA, AADE, JDRF, ACP on the adult side, and AAP put together a transition toolkit, and in that toolkit is a skill set. It’s available online. We have to get them ready to go off on your own. We talk about how to learn how to cook, how to fill prescriptions. It’s been extremely valuable. I also have to ask: your slides showed that the less often that patients bolus, the higher their A1cs. Yesterday we talked the lack of utilizing executive function. They don’t need help with their basal, but its like after age 10 they have no memory. What tools do you use to get your patients to bolus?

Dr. Phillip: Technology; it can be a way to remind them like using a cell phone, it can be through SMS messages that they send whatever measurement to the parents. There are plenty of ways that depend on available technologies.

Dr. Klingensmith: But they only work for a short period of time unless the pump is going to do it for them. We have published that.

Dr. Rewers: There is no answer. The panel is hopeless [laughter].

Dr. Klingensmith: One more thing – consequences. Parents are very reluctant to have consequences for teens not doing what they need to do for their diabetes. They have to have rewards and consequences for diabetes care but that is a real battle.

Q: I work within a school district in Colorado. When you were talking about transition programs, we’ve seen that they have started to move back further and further, and now legally it needs to happen starting in sixth grade.

Dr. Klingensmith: We are starting a middle school transition program.

Q: Are there legal requirements?

Dr. Klingensmith: No, and actually middle school parents have been asking for this for a long time.

Q: And they saw an improvement? Do you have the adult doctors come to your pediatric visits? If so, when does that start?

Dr. Klingensmith: No we don't.

Comment: I just want to pick up on the question of how you get patients to bolus. We just did an intervention with colleagues at Cincinnati Children’s Hospital. It’s important to have a collaboration between the educator, provider, and child. We start by asking patients what makes diabetes hard for them. We’ve had kids start to cry, and say that nobody ever asked what makes it hard for them. And so I think behavioral skills training is very valuable.

Dr. Rewers: I’ve been trying to positively reinforce this kid for the past three years, but he or she is losing ground. Short of finding a foster family, what can you do?

Dr. Maahs: I think learning a little bit about motivational interviewing. I just try to ask a child, “What can I do to help you today?” They know they are supposed to check their blood glucose more and bolus more. Des made an important point of being encouraging. I will ask them, “Will it do any good if I get angry at you?” They almost always say, “No, no that is not gong to help.”

Dr. Dunger: I’m on the touchy feely end I think. You need to find out what other pressures exist, and try to come to some negotiation about what sort of change they can make. You might need to keep seeing them over and over until you can negotiate a change, or maybe a change will be quick. Don’t underestimate the importance of keeping in touch — patients do change. I had a patient who had fifteen admissions with DKA, and then three years later her life improved dramatically. I don’t understand what changed, but it did. Keeping in touch was our only achievement.

Dr. Klingensmith: I pass.

Dr. Phillip: We call that the 10 o’clock. It is a code. We know that we are going to take several steps. The basic idea is tailoring the therapy to each patient. We have a team meeting and discuss what to do with this specific patient. We increase the frequency of their visits to very month. We talk to the patient about ways to reward the kid for improving – depends on the age of the child of course. We almost always put them onto a blinded CGM and discuss the results, and sometimes it works, but not always.

Dr. Klingensmith: Why blinded?

Dr. Phillip: It increases their understanding. Blinded you can calibrate at the end when you download it so it does not require as much work on their part. 

Dr. Klingensmith: I think that an A1c over 10% and you have done everything you can do, then there is some barrier we don't know about and finding that barrier might be very difficult. Social workers and psychs might be better equip to find that barrier. Often in my experience that barrier is that someone is abusive in the home.

Q: We always say that the reason we have these meetings is because we’re worried about complications with kids, but we never talk about complications with these patients. Since helping patients avoid complications motivates us to work so hard, why don’t we help them find that same motivation? Couldn’t talking with patients about complications be a positive thing?

Dr. Maahs: I think you are right. I think we do need to education in our patients even the teens on the risk of an elevated A1c and future complications. I don't know if hammering on that too much motivates them. Dr. Polonsky put up a slide “True or False? Diabetes is the Leading Cause of Blindness.” The next slide said “Uncontrolled diabetes is the leading cause.” I think that the important thing is to empower patients to know that they can control this. You need to ask them what they are worried about because different families have different concerns. 


Session: Case Discussion and Asian Perspective

Panel Discussion

Linong Ji, MD (Peking University People’s Hospital, Beijing, China); Richard Bergenstal, MD (International Diabetes Center at Park Nicollet, Minneapolis, MN); Satish Garg, MD (University of Colorado, Aurora, CO); Robert Rizza, MD (Mayo Clinic, Rochester, MN)

Q: Do you find that in Asian patients, you use different add-on therapeutic agents? Also, do Asian patients fail therapy earlier?

A: There are few long-term studies in China. For example, when we put a patient on an SFU, we see the same Nike curve as was seen in ADOPT or UKPDS. So I think the response and treatment profile is more or less the same.

Q: I was struck by how big that drop in A1c was seen in your case. I know you showed data about DPP-4 inhibitors being more effective, but wow — a big drop in A1c and less insulin. Was the acarbose causing a GI upset, or were there other adherence or compliance issues that have now been resolved?

A: We run many clinical trials with DPP-4 inhibitors. My physicians always tell me that with some patients, they work extremely well. If you look at the comparison study between acarbose and DPP-4 inhibitors, they have the same average efficacy. But in individual patients, there could be a big difference.

Comment: I work at a prison in Kansas. I’m not that familiar with DPP-4s. What are the brand names?

Dr. Garg: Janumet, Onglyza, Tradjenta.

Comment: Yes, we don’t have those.

Dr. Rizza: Don’t use Glyburide. It’s an evil drug.

Comment: That’s all we have.

Dr. Rizza: That’s the only thing the entire room will agree on. Patients using it have a high frequency of hypoglycemia. Convince your formulary to use glipizide. It’s generic.

Dr. Garg: Yes, it’s only $4 per month.

Q: I have a patient who was well controlled on metformin and lost control, should you add insulin? Can you drop the metformin or should you stick with both?

Dr. Rizza: I work at a strange place. The common approach is to add on sulfonylurea, but there is no common approach at Mayo. Usually you add a little bit of sulfonylurea and see a trend upward. A common mistake that people tend to make is to take sulfonylurea in the morning.

Q: So you wouldn’t do insulin but sulfonylurea?

Dr. Rizza: Some patients would say, “Sure, I’ll take a shot of insulin.” But yes, the most common approach would be to add on sulfonylurea. Also, it’s generic and it’s cheap. That being said, if we knew the best method for sure, then all of this would be very straightforward.

Comment: Some believe that sulfonylureas are hard on the pancreas by forcing the beta cells to produce more insulin.

Dr. Rizza: There is little if any data that sulfonylureas decrease insulin secretion. There are a variety of reasons everyone should be concerned. I will show you data that shows that something happens to beta cells when you increase sulfonylureas and increase insulin secretion. Pioglitazone is a potent stimulator of insulin secretion, and it acts on the beta cell. Even looking at UKPDS, people ask about the data. The reason you should not use sulfonylureas, however, is to not get hypoglycemia, not because of glycemic dependence or beta cell failure.

Dr. Garg: Try DPP-4s once they are generic.

Dr. Rizza: I will challenge you on looking at the data on DPP-4 vs. sulfonylureas. There are reasons why we keep having these conversations. It is hard to show the complete advantages of any class.

Dr. Garg: I still think that sulfonylurea will have a very important role. I still use them a lot more than DPP-4s because people cannot afford DPP-4s. I would rather put my patients on a sensor and look at their profile see how they are doing. That is a much better option than using these expensive drugs.

Q: I have not used a sulfonylurea on my patients in the last 10 years, and I don’t know why people are using them.

Dr. Rizza: I work at the Mayo Clinic, and I treat a very large number of people with diabetes. We use sulfonylureas all the time. If you look very carefully, the drug with highest mortality – epidemiology – is insulin. The sickest people get put on insulin. If you could read a crystal ball, you would put your patients on metformin. After that, EDIC says that patients on insulin and sulfonylurea have lower mortality than the control group. It is very hard to impune sulfonylurea, except for weight gain and hypoglycemia. The marker is how many prescriptions are being written, which is a huge amount for sulfonylureas. The amount of eye disease, renal disease, cardiovascular disease are all falling despite the high use of sulfonylureas.

Dr. Garg: I use sulfonylureas a lot in people with type 2 diabetes. Sulfonylureas have been thrashed partly because of the pharmaceutical companies, however, I don’t think this is a pathway we should follow. I think sulfonylureas are very effective, you just have to make sure hypoglycemic events can be managed. The weight gain is only 1-2 kilo difference. I don’t think the drugs need to be bashed the way they have been. None of these new drugs have long-term data.

Dr. Bergenstal: There has hardly been a head to head.

Dr. Garg: I think that the GRADE study will tell us.

Dr. Bergenstal: Companies don’t want to include sulfonylurea because it will look too good. We are waiting for the CAROLINA and GRADE studies.


Meet the Peers – Adult

Panel Discussion

Sir George Alberti, MD (Imperial College, London, UK); Richard Bergenstal, MD (International Diabetes Center, Park Nicollet, Minneapolis, MN); Hans DeVries, MD (University of Amsterdam, Netherlands); Irl Hirsch, MD (University of Washington, Seattle, WA); Stephanie Amiel, MD (Kings College School of Medicine, London, UK); Dr. Satish Garg (Barbara Davis Center, Aurora, CO); Abhimanyu Garg, MD (UT South Western Med. Center, Dallas, TX); Angelyn Bethel, MD (University of Oxford, UK); Andy Karter, MD (Kaiser Permanente, Oakland, CA)

Q: What about TZDs? There are type 2 diabetes requiring hundreds of units of insulin who are not well-controlled.

Dr. DeVries: I have always had a problem with TZDs – why use a medication that increase body weight?

Dr. Alberti: Has anyone tried using GLP-1 analogs?

Dr. Amiel: We start with metformin added to the insulin. You can try a TZD, but it causes weight gain. We restrict to metformin. We do use incretin analogs, and U-500, but we tend to use it in a pump. Or you send patients to the surgeon.

Dr. Hirsch: As far as GLP-1, we retrospectively went back and looked back at patients who were on 200 units of insulin per day and who were not controlled. We looked at 13 patients after 12 months in an uncontrolled trial. Their weight down, their insulin dose down, with most already on U-500. There is something about the GLP-1. We have seen a beneficial effect. As far as insulin is concerned, patients were on 200 or 300 units per day. This makes a situation more simple than it is. When I see a patient like this, the first thing we usually see is a patient with type 2 diabetes on 100-200 units per day of glargine. What is the benefit of that? The whole point of insulin analogs is to have a reduction in hypoglycemia. Usually their A1cs are 9%-10%. I actually believe, for basal insulin, NPH insulin does a better job. The peak is unpredictable, but when I give it at bedtime, I can start with a lower fasting glucose. I get the fasting glucose down near normal, and they have a better chance of doing better the whole day. We published a case study when a patient saw a reduction in A1c from 9% to 6.5% by switching from glargine to NPH at bedtime. With long-acting analogs, I will give it in two different depots. This gives a broader surface area with better insulin kinetics. I can actually not use U-500 a good amount of the time.

Dr. Amiel: I have the same point about splitting the dose. Once the patient is over 40 units per dose, I usually split it. I find that Levemir is not good in particularly large patients.

Dr. Alberti: It is sad that we’re having this discussion about failure.

Dr. Garg: This comment is related to adding treatment. Is there any data to show that SMBG may be useful in people not treated with insulin therapy?

Dr. Karter: The talk that I gave today was essentially about that group. There is definitely evidence to suggest that SMBG has a real educational and motivational benefit in patients not on insulin.

Dr. Garg: What frequency do you use SMBG in patients not using insulin?

Dr. Karter: I am not sure. The Cochrane report showed, consistent with out study, that starting patients on SMBG actually lowered A1c, even patients that are diet controlled. However, that type of benefit waned over time. We found that with a one strip change there was about a 0.15% change in A1c. With a decrease in strip use among prevalent users of SMBG, there was a 0.15% increase in A1cs in pharmacologically treated patients. In our studies, we restrict it to staying on stable therapy. It gets complicated to sort out.

Dr. Alberti: We need a reality check on this. Let’s do a calculation. If all type 2 diabetes patients in the UK were monitoring four times per day, it would cost £12-15 billion a year. That is more than 10% of our health service budget.

Dr. Karter: I am not recommending chronic use of SMBG. I think we use it too much. We need to be much more selective. I like the STeP approach – occasional use in a structured way.

Dr. Bethel: The take home message is that SMBG is a way to increase self awareness and self care. It should be use in an intermittent fashion. The only thing that could really be impactful is behavior. That might be the more reasonable teaching point.

Dr. Karter: That is what I was trying to say. Exactly. That is what we are seeing based on our data. I’m not a clinician.

Dr. Alberti: But you are a peer.

Dr. Amiel: In the study you showed this morning, the first time data with SMBG in patients with type 2 diabetes showed improvement in glycemic control. Your study looked at regular home SMBG with some feedback against CGM. The next study should look at whether the benefit of checking the blood sugar four times daily can be achieved with less frequent monitoring.

Dr. Bergenstal: Yes. For me, we need patients and clinicians checking intermittent blood glucose levels together. I am not convinced that teaching someone principles and setting them loose is the right way. There needs to be some feedback.

Dr. Satish Garg: What is the role of sulfonylurea in the management of type 2 diabetes?

Dr. DeVries: I think they are put in the wrong parts. No one defends them. ADVANCE is a very powerful study to illustrate that their safety record is great. They do have downsides – increases in hypoglycemia and increases in body weight. They are extremely cheap, which in a large part of the world is important. I think they tend to be underrated these days. This is not the perfect class of medication, but…

Dr. Amiel: I completely agree. They have been around a long time and we know exactly what they are. Every time I open up the newspaper, I see new problems with new drugs. Dr. Gerstein presented data this morning that they have no cardiovascular disease disadvantage. I think they still have a lot to offer.

Dr. Alberti: I would emphasize the lack of cost. I did a lot of work with colleagues in sub-Saharan Africa. The first drug of choice was glibenclamide. We could treat someone with that for less than $5 per year. The only caveat, which we found in the UK as well, is hypoglycemia in older patients.

Dr. Garg: Is there any data from Kaiser on this?

Karter: It is used quite a bit. We do see hypoglycemia. But we see hypoglycemia in every dryg, including metformin. Hypoglycemia is something you have to be really careful with. Especially in high risk patients like elderly and low GFR. We use it. The issue of cost is very important.

Bethel: Not much to add, except that they have been painted in a bad light for a long time. They are good drugs. They are effective drugs. They do exactly what they say on the tin – lower bG. Like every other diabetes drug, they have a problem – hypoglycemia. It’s not about not using it; it’s about being aware. They are fantastic drugs, and I don’t think there is any good reason to avoid them.

Dr. Alberti: The shadow of tolbutamide being cardiotoxic – it’s taken a long time for that to disappear.

Dr. Hirsch: It’s very interesting to me. Sulfonylureas – if you look at the AACE guidelines, I’m an author on them. But I have a problem – I was not involved in that part. I was in the insulin section. Sulfonylureas were not painted in a good light for the AACE guidelines. Many endocrinologists in the US feel very strongly about that. I think we know what we have in terms of the bad and the good. The real issue is their duration of effectiveness is not going to be very long. The ADOPT study showed that quite well.

Dr. Bergenstal: Before GRADE, CAROLINA comes out. That’s comparing a sulfonylurea to a DPP-4 in a long-term study.

Dr. Amiel: In the elderly, I tend of avoid glibenclamide because it is long acting. The problem we have is that it takes us so long to change everything. I think it’s perfectly fine to add a sulfonylurea to metformin, provided you then move when A1c is no longer well controlled.

Q: About sulfonylureas, what about the progressive nature of type 2 diabetes and destruction of beta cells? All those Fonseca articles that sulfonylureas may take up more of the beta cell use?

Dr. Bethel: I think that’s a nice idea – it seems like it is pushing on the beta cell to make insulin, to make insulin, but in reality, we have never seen any data from any randomized controlled trial that there is a difference with sulfonylurea accelerating beta cell destruction.

Dr. DeVries: If you look at the UKPDS data that everyone knows, beta cell function declines linearly over time. But the graph stops a few years after diagnosis. It’s a linear slope going down, and it’s extremely tempting to extrapolate and say that after 10 years of diabetes, beta cell function is zero. But our emergency departments are not overflowing with people with long-standing type 2 diabetes and DKA. I think this has to level off.

Dr. Alberti: Anyone know of any studies showing total insulin secretory failure?

Dr. Garg: ADOPT was biased against sulfonylureas. If you give sulfonylureas to a patient with an A1c of 10%, you won’t see hypoglycemia as much as you would see if they have an A1c of 7.5%. My take on ADOPT is that it shows sulfonylurea are associated with more hypoglycemia – but in what setting? Very early diabetes.

Dr. Bergenstal: Although GLP-1 are complicated, it’s amazing they can have an effect 15-20 years into diabetes. there is still something there.

Dr. Amiel: I don’t want to undo the reassuring message. At ADA, Dr. Ralph DeFronzo presented two-year outcome data comparing use of metformin and then adding a sulfonylurea when patients were losing control. The comparator was starting exenatide, pioglitazone, and metformin at diagnosis. Everyone’s A1c went down. His triple therapy patients stayed stable over two years. The other group started to drift up.

Dr. Bergenstal: They didn’t drift up quite the way we’re used to seeing. They ended up at 6.6%.

Dr. Amiel: The drift up was mild in the sulfonylurea group.

Dr. Hirsch: I think that there’s a mixed message on long-term sulfonylurea use. I don’t think we understand the beta cell biology as much as people think we do in long-term type 2 diabetes.

Q: What are you doing for prediabetes patients in Europe? Here, usually diet and exercise doesn't work for us. What do you start them with as far as medication is concerned?

Dr. Bethel: We do exactly what you do here. We preach diet and exercise. We do everything we can to get them involved. Evidently, my preaching isn’t good. If I can get them in a community program like Weight Watchers, that tends to work. We are not stepping to medication as quickly as our US counterparts.

Q: In the UK, you said you had only one patient with prediabetes on metformin?

Dr. Amiel: There is evidence that metformin can be effective. If I were going to use a medication, it would be metformin. I think it’s coming.

Dr. Garg: Several studies have been done. Can we prevent progression from IGT to diabetes? The problem is when you treat them with a sulfonylurea or metformin or a TZD, what you see is the treatment group does much better. The critical question is prevention. When you stop, what happens? There is very little evidence that follows patients over time and looks at the outcome.

Dr. Alberti: Recently, I was peripherally involved in a study in Chennai, India. People were screened for prediabetes. If they were found to have it, they were given mobile phones. After an initial chat, they just got text messages every couple weeks. Over the course of two years, it halved the number of people who got diabetes. That was just based on diet and exercise and behavioral changes. We need to emphasize the importance of that, without any metformin or drugs.

Comment: Canagliflozin looked quite interesting – they recently showed 104-week results with less hypoglycemia vs. glipizide, less weight gain, and a better A1c at two years. If prevents diabetes, maybe we could look at that area.


Industry Sponsored Dinner (Sponsored by Novo Nordisk)

GLP-1 In Focus

Neil Goldberg, MD (Cedars-Sinai Medical System, Beverly Hills, CA)

In the introduction to this dinner presentation, Dr. Neil Goldberg relayed that Novo Nordisk’s main insulin production facility is fully run by wind power – we continually applaud the company’s focus on sustainability. Next, Dr. Goldberg took the stage to discuss GLP-1 agonists. He started with the ADA/EASD position statement, noting that GLP-1 agonists are the only agents on the list that lead to weight loss, have high efficacy, and have a low hypoglycemia risk. He highlighted familiar facts about the drug class, namely that these agents stimulate insulin secretion and inhibit glucagon secretion in a glucose-dependent manner. He emphasized that fear of injection should not be a barrier to treatment, stating that providers should not be needle-phobic in their prescription patterns. Turning specifically to Victoza (liraglutide), Novo Nordisk’s GLP-1 agonist, Dr. Goldberg addressed concerns about pancreatitis by noting that diabetes itself leads to an increased incidence of pancreatitis, and furthermore that postmarketing data does not seem to confirm the association. He downplayed concerns about nausea as a side effect, arguing that it is part of the drug’s intended effect, and usually subsides over time.

Questions and Answers

Q: You mentioned gastric bypass surgery and its effect of elevating GLP-1. Does that effect last? I was wondering if GLP-1 agonists could be an appropriate therapy for patients after the effects of gastric bypass begin to wane.

A: There have been some cases of using Victoza after gastric bypass, but it’s an off-label use, and contraindicated from that point of view.

Q: If there is a physiologic defect in type 2 diabetes that they are deficient in GLP-1, why not just give a small amount with each meal – cover it in a physiologic range?

A: That is a great question. Exenatide [BMS/AZ’s Byetta] does that. However, patients experience fewer side effects with a more steady amount of GLP-1. I have done this with my best patients – one of the drugs works better post-prandially and the other works better basally. There are differences. If the prandial use of GIP did not cause significant GI problems, it would be a great approach. What about putting it into a pump? What a great idea would that be, but nobody has done that. You have seen patients on Victoza whose post prandials are not well controlled and then you need to add something to the mix. That is why these are not cures. They are not perfect.

Q: If it is physiologically decreasing insulin and insulin controls glucagon, why not use insulin instead of GLPs?

A: You can. We do that in people with type 1 diabetes. Glucagon is elevated in type 1s also. But what else does insulin do? It causes hypoglycemia and it causes weight gain.

Q: Isn’t insulin physiologically better?

A: It is if you inject it into the portal vein. We don't do it physiologically; it has to go through your whole body. The concentration we use subcutaneously is about 10 times what we could use if we injected it straight into your liver. Look at transplants, where are the cells put? In the liver, because that is where the insulin has the greatest impact. Insulin is put up as the gold standard; you can do anything with insulin but it has side effects. Your point is very well made about the physiology.

Q: The other thing I saw on the package insert is that it wasn’t approved as a first-line therapy, but now AACE guidelines say it can be used as a first-line therapy

A: I think that this was one of the FDA fights — they don’t want to encourage starting with it. But it’s your practice; you can do what you want. The official studies have been done saying that it should follow metformin, but there have been monotherapy studies showing that it works.

Comment (Novo Nordisk representative): It’s not a recommended first-line option, but in the event that metformin is not an appropriate choice, then it is approved as a first-line choice. But we are not authorized to promote it as a general first-line therapy.

Q: If it physiologically decreases insulin, and insulin controls glucagon, why not use insulin instead of GLPs?

A: You can. We do that in people with type 1 diabetes. Glucagon is elevated in type 1s also. But what else does insulin do? It causes hypoglycemia and it causes weight gain.

Q: Isn’t insulin physiologically better?

A: It is if you inject it into the portal vein. We don't do it physiologically; it has to go through your whole body. The concentration we use subcutaneously is about 10 times what we could use if we injected it straight into your liver. Look at transplants, where are the cells put? In the liver, because that is where the insulin has the greatest impact. Insulin is held up as the gold standard; you can do anything with insulin but it has side effects. Your point is very well made about the physiology.

Q: Can you take Victoza at any time? If so, is there any change in appetite?

A: Yes you can give it any time, which is counterintuitive because historically we’ve always been told to take medicines before you eat. But a drug like this can be taken at any time, although you should try and take it at the same time each day. It’s a very friendly drug, and the effect is the same whenever you take it. Regarding the appetite issue: ask your patients about it, don’t wait for them to approach you about it. Ask if they can still eat a full meal anymore. Patients become empowered; people are in control of things they couldn’t control before. When it comes to people who have had diabetes for 20 years, don’t you think they would have lost weight if they could? This is a pathology.

Before we end, I’m going to share a theory with you. We don’t really know exactly what diabetes is. I was looking at these slides earlier, and I realized something. If you think about the problems associated with diabetes, what do they all have in common? I think all the symptoms of diabetes are there to ensure appropriate carbohydrate nutrition to the brain. Maybe insulin deficiency is due to a sensing mechanism in the brain — maybe the brain doesn’t recognize the sugar, and wants more. The action of glucagon, gastric emptying, muscle and fat cells — they can all raise sugar levels. The commonality of all of these things is that the brain needs something. Maybe we’re looking at the wrong place. All our medicines may fix the signals, but are not curing the disease. I throw this out there for you to think about; there’s no proof behind what I suggest as of yet.


Keynote Speaker

Non-Profit and Industry Collaboration: JDRF Perspective

Richard Insel, MD (JDRF, New York, NY)

Dr. Insel began his talk with the JDRF’s three goals: to cure, treat, and prevent type 1 diabetes. He emphasized that partnerships are key to this mission. The JDRF has worked closely with government research and regulatory agencies, holding workshops and pushing objectives such as the artificial pancreas (see page 32 of our Keystone 2012 report at for Dr. Insel’s talk on JDRF-FDA collaboration). Partnerships with industry provide other fruitful opportunities for cooperation, as non-profits such as the JDRF can help bridge the funding gap (“valley of death”) between basic science research and clinical development. Dr. Insel noted that when JDRF cooperates with industry, it is not investing in companies; rather, it is investing in specific therapies. He highlighted the importance of addressing gaps in research, citing two JDRF partnerships to help develop a more stable, pumpable glucagon for bi-hormonal closed loop artificial pancreas systems as examples of such efforts. The organization generally asks for at minimum financial matching from the partner company, and holds them accountable over the course of the collaboration with milestone-based payments. It engages in three primary categories of partnership to help companies internalize and externalize type 1 diabetes research and objectives: co-funding with industry of early-stage biotech projects or research in academia, direct support of internal company projects, and non-financial strategic partnerships with industry. Dr. Insel also highlighted the increased opportunities for cooperation with the pharmaceutical industry in the precompetitive space, in areas such as biomarkers and imaging.

  • Dr. Insel began his presentation with the JDRF’s three goals: to cure, treat, and prevent type 1 diabetes. Dr. Insel noted that partnerships can accelerate the process of getting drugs and devices into the hands of patients. He briefly discussed JDRF’s work with payers and healthcare systems to help improve reimbursement for CGM. He also highlighted the organization’s work with the FDA, including getting the artificial pancreas placed on the agency’s Critical Path list in 2006 and delivering regulatory guidance for artificial pancreas systems.
  • Today’s focus, however, was on cooperation between non-profits and industry. In Dr. Insel’s mind, such partnerships are fundamental in bridging the financing gap (or, as he put it, “valley of death”) between basic research and clinical development. JDRF’s goals in its industry partnerships are to advance research translation and product development, especially in “gap areas” such as the development of a stable, pumpable glucagon. Dr. Insel was clear that these partnerships did not represent an investment in a company, but rather an investment in a specific candidate therapy. The projects are contracts, not grants, involving at minimum  matching funds from the company, milestone-driven payments, a capped royalty stream, and high accountability. JDRF generally maintains walk-in rights should the partner halt the development of the candidate.
  • Dr. Insel listed three major categories of JDRF-industry partnerships. In the first, JDRF and a pharmaceutical company co-fund an innovative, high-risk project at an academic institution or biotech company with translational potential. J&J, Sanofi, Lilly, and Pfizer have all been partners on these types of projects. In a variation on this theme, JDRF also co-funds precompetitive consortiums to support key development areas such as biomarkers or imaging and has partnered with IMIDIA of the Innovative Medicines Initiative (IMI). In the second partnership category, Industry Discovery & Development Partnerships (IDDPs), JDRF co-funds a specific internal company project. IDDP partners have included ViaCyte, Amylin, BD, NGM Biopharmaceuticals, Animas, Medtronic, and KalVista. For the last major category, JDRF contributes strategy and expertise (but no funds) to help internalize and externalize type 1 diabetes initiatives. Novo Nordisk has benefitted from this type of partnership.

Questions and Answers

Dr. Peter Chase (Barbara Davis Center, Aurora, CO): I would want to make sure people know that the artificial pancreas wouldn’t be where it is without the JDRF. It’s not just the study they conducted, but also the grants they’ve put out. We’re working with them on a number of areas.


Predictions on How the Insurance Industry Will Respond to Healthcare Reform

Jack Lord, MD (University of Miami, FL)

Dr. Jack Lord (University of Miami, FL) gave a focused talk on healthcare reform in general and the Affordable Care Act (ACA), a law that he characterized as “not real healthcare reform” – rather, he views it as an “ideological expression of the frustrations with healthcare.” Dr. Lord sees the ACA having both positive and negative impacts on diabetes care: better access to insurance and more prevention/coordination of care, with the caveats of cost pressures on providers/therapeutics/devices, issues with access to providers, and reductions in discretionary spending that further complicate the regulatory approval landscape. Dr. Lord does not think Accountable Care Organizations (ACOs) will have a near-term impact, and in the long-term, thinks they will have a “minimal and isolated impact.” That said, he does believe bundled payments will be a part of the future healthcare landscape – we agree. He had three recommendations for industry to succeed in this new era: 1) rigorously assess the cost-effectiveness of products (“clinical effectiveness is necessary but not sufficient”); 2) acquire new skills in medical economics, predictive modeling, and consumer insights; and 3) “Data, data, data.” In concluding, Dr. Lord called for “real healthcare reform,” which “starts with people like yourselves as you think about different ways to provide care.” Given his background at Humana, we appreciated his closing remark: “insurers, payers, and the government will always adjust to win. The health insurance business is just like the casino business – the house sets the rules, and the house always wins.” 

  • “How do insurers think?” Dr. Lord identified a few answers: 1) cautiously (a paradox, since the entities that hold risk are very risk averse); 2) very attuned to practitioner attitudes and preferences (e.g., KOLs, academics, and guidelines re critical to coverage decisions); 3) focused on the return on investment for any new initiative (increases the scrutiny of medical economics of every new therapy); and 4) no free lunch (mandates in one domain will be offset by reductions in other areas).
  • Dr. Lord expressed concern about how the Affordable Care Act (ACA) might increase the cost of insurance premiums. The risk with ACA is that only sick people opt in to get insurance. Dr. Lord wondered how much higher insurance premiums will go to pay for the mandate and changes in the structure of insurance. From an actuarial perspective, he also highlighted the lag of 20-32 months as insurers look backwards to calculate prices for the future. Unfortunately, the ACA causes insurance companies to add a “fudge factor” for uncertainty. This drives up prices and has potential to worsen the risk pool and lead to a death spiral.
  • Regarding Accountable Care Organizations, Dr. Lord was negative on their potential to change much of anything. He noted, “we have been here before” – HMOs in the 1980s and PSOs in the 1990s. All these programs have a common theme of organization, sharing risk, and coordination of care. While these are indeed “very laudable goals,” they were never achieved on a structural basis. Part of the problem is that people were collecting money up front, but they did not understand the risk, and thus didn’t have money to pay for care. The fear for ACOs is that they will appeal to sick people only, which will drive up the cost of services over time. Overall, Dr. Lord does not see ACOs having a near-term impact, and in the long-term, he thinks there will be few areas of success (“minimal and isolated impact,” “issues of talent and risk tolerance”).
  • “Bundled payments will be a part of the future landscape. Providers and organizations will have to figure out how to share and where to share.” In Dr. Lord’s experience, “sharing is not a strong talent.” He believes providers will find themselves at some level of risk and there will be an incentive for better outcomes (e.g., you cannot participate in Medicare/Medicaid if you don’t demonstrate outcomes). For industry, Dr. Lord expects a “relentless pressure” on the prices for therapies and devices. He sees this not only at the unit level (e.g., test strips), but in overall expectations that therapies and devices lead to better care, better outcomes, and lower costs.


Public Policy Issues and the ADA’s Role

Robert Ratner, MD (American Diabetes Association, Alexandria, VA)

Dr. Robert Ratner eschewed his slides in favor of candidly speaking to the audience, giving everyone a chance to hear an excellent overview of the ADA’s public policy initiatives: research funding (including the Special Diabetes Program), healthcare reform, discrimination, active involvement at the FDA through open public hearing presentations, establishment of guidelines and standards of care, getting representatives on important government committees (Medicare Payment Advisory Committee, the US Preventative Service Task Force, PCORI), targeted research based on unanswered questions from special conferences (e.g., the ADA Weight Loss and Weight Regain Symposium), the new Pathway Program to fund young and innovative researchers, and CME sessions (already offered for PCPs, and notably, ADA will provide these to medical schools, academic hospitals, nursing schools, and graduate programs free of charge). He emphasized that the ADA’s policy initiatives focus on its single constituency: people with diabetes and their families and caregivers.

  • Research funding: “Every year, we are pushing to increase the overall budgets at NIH, AHRQ, and CDC to promote increased research within the federal government.” Dr. Ratner emphasized that ADA is not pushing exclusively for diabetes research; rather, the Association is hoping to benefit the entire pie and grow it (“we don’t want to get into a match with arthritis or cancer”). One diabetes-specific research funding area ADA advocates for is the Special Diabetes Program, which must be renewed annually. The Program’s approximately $300 million in funding per year is split evenly between type 1 diabetes research and the Indian Health Service (the latter is to correct disparities, as this population has an overwhelming prevalence of type 2 diabetes).
  • Health reform: “ADA from the beginning supported the Affordable Care Act.” Dr. Ratner highlighted two key benefits of ACA for people with diabetes: increased access to care for people with preexisting conditions and essential benefits package that must be provided by the exchanges. Factors: three tiers of insurance that must be provided by all of the exchanges. ADA is now advocating at the state level to ensure minimal essential benefits packages include coverage for things like insulin, anti-diabetic therapies, and blood glucose meters.
  • Discrimination: “For way too long, patients with diabetes have been discriminated against.” ADA has advocated on behalf of commercial truck drivers, who were not legally allowed to drive if they were on insulin therapy. This resulted in poorly controlled patients on orals refusing to go on insulin, or such patients getting on insulin and losing their jobs. ADA has also focused on discrimination against airline pilots, FBI employees, and many other occupations.
    • The Association’s “Safe at Schools” program has also been a focus in recent years. Dr. Ratner mentioned the recent California Supreme Court hearing, where the state’s nursing association argued that only a licensed nurse can give insulin and glucagon injections. “That means a 17-year-old child cannot take an insulin shot at school. That’s absurd. That’s absolutely absurd.” The ADA has fought in many states in favor of allowing patients and non-nurse caregivers to administer insulin in schools. ADA is very optimistic they will win the case, and Dr. Ratner hopes it sets the stage for multiple states [Editor’s note: the CA Supreme Court recently voted in favor of allowing non-nurse, unlicensed public school personnel administer insulin to children with diabetes. Read our full report at].
  • Involvement at the FDA: “Historically, we have not had any impact [at the FDA]. We have not spoken or discussed with the FDA. That’s no longer the case. Dr. Ratner mentioned that over the last year, every single FDA hearing that concerned diabetes (devices, ophthalmology/DME, or even most recent hearing on rosiglitazone) has had an ADA representative in the open public hearing. Said Dr. Ratner, “During the public hearings, we advocate for people with diabetes. We do not talk about whether this drug is better than that drug. We talk about the unmet needs for people with diabetes. Do we have adequate therapies? Do we need more? The premise is from here on out, we will be at the FDA as an active participant giving views of what is needed.” We think this is awesome and salute the ADA for getting involved.
  • Establishment of guidelines and standards of care. ADA’s professional practice committee follows the Institute of Medicine’s guidelines on guidelines documents. ADA updates their practice guidelines annually to keep things up to date (“Every January, you know that you will have new standards of care from the ADA”). The focus of these guidelines will be more and more on the individual patient. Dr. Ratner also emphasized that ADA “will never accept any funding for the development of our guidelines.”
  • Representatives on important government committees: Dr. Ratner mentioned EMDAC patient representative Rebecca Killion, calling her “the single smartest person on that advisory committee – a woman with diabetes. She thinks logically as opposed to many other people.” ADA nominated Dr. Bill Herman from the University of Michigan to be a member of the influential Medicare Payment Advisory Committee (MEDPAC). ADA is also trying to get a representative on the US Preventative Service Task Force (the organization identifies category A and B criteria that have to be paid for without a copay in the Affordable Care Act). Last, ADA is working hard to get someone on the Patient Centered Outcomes Research Institute (PCORI) – the organization has $300-500 million dollars per year for patient centered research.
  • ADA is now conducting targeted research guided by unmet needs identified at special conferences like the Biologic Mechanisms of Weight Loss and Weight Regain Symposium (see our report at Said Dr. Ratner, “With all the conferences from here on out, we’re trying to find unmet needs and identify a targeted research program.” Indeed, Eli Lilly has funded a targeted research program to answer questions raised from last year’s conference on diabetes in the elderly, and Dr. Ratner said a program is also expected to stem from the weight loss and regain symposium.
  • ADA’s new pathway program will fund the next generation of diabetes researchers. Said Dr. Ratner, “Unlike the typical application, we’re asking to find the most brilliant people who are capable of thinking outside the box.” The five-year funding of $1.6 million per person is aimed at “trying to change the paradigm of research in academic institutions.”
  • “We have to recognize that there are never enough diabetes specialists to care for all the people with diabetes. It has to be done at the level of primary care. It is incumbent on us to make sure they know the most up-to-date information. And we’re committed to doing that.” ADA just had its second annual Diabetes is Primary course (“essentially Diabetes 201”). This one-day program takes practicing PCPs and educates them on what’s changed in diabetes. In Philadelphia at ADA 2012, the first course had 130 attendees. It’s also on the web as a CME program, and 3,000 people have taken it for credit. We think this is an outstanding move on ADA’s part and we hope it can spread and grow – the needs for PCPs to be well educated on diabetes is profound.
  • ADA is initiating the Diabetes Academy, a CME program based upon standards of care, position statements, consensus conferences, and research conferences. The up-to-date, vetted material will be offered to medical schools, academic hospitals, nursing schools, and graduate programs at no charge. Said Dr. Ratner, “We can get the best speakers, the people who are integrally involved in the development of the data.”


Role of Academia in Achieving Targets in Diabetes Care

Robert Rizza, MD (Mayo Clinic, Rochester, MN)

Dr. Robert Rizza emphasized that the role of scientists in academia is to recognize and articulate issues that need to be addressed in order to improve the care of patients with diabetes. In his view, researchers are there to provide the scientific underpinnings for improved safety and treatments for patients. He explained that academia has provided scientific evidence for methods of diabetes prevention and reversing complications. For example, academia has helped to develop models that provide plausible mechanisms by which diabetes alters glucose metabolism. Turning to how academia fits into the whole healthcare system, Dr. Rizza highlighted the importance of community in academia, calling the group a “we.” In particular, he asserted that academia should work with both clinical and industry partners – we certainly agree. In fact, Dr. Rizza thinks that such collective effort is the only way that academia will survive and continue to provide necessary and critical services to patients. Ending his presentation, Dr. Rizza called on academics to serve as patient advocates – given their less biased position, those in academia are in a good position to ensure that patients have access to better healthcare and treatment.

  • Academia has been a leader in “mitigating risk and maximizing benefits” of new therapies. Dr. Rizza emphasized that animal models have been invaluable in hypothesizing the mechanisms of glucose counter regulation in humans. Academia has also been involved in the development of pancreas transplantation and islet transplantation and has led efforts in developing closed-loop insulin infusion systems. Animal models developed in academic centers have also helped establish knowledge of glucose regulation mechanisms, which helps to avoid patient risks. Dr. Rizza believes that academia will continue to lead in this area.
  • Dr. Rizza highlighted that he led and served on the ADA committee that established the standard of care for people with diabetes as an A1c of less than 7% (in the absence of hypoglycemia and other comorbidities). We hope to see such committees formed in the coming years to develop standard of care measures for time in range! Dr. Rizza also discussed his leadership and service on multiple other groups and panels, including AACE, the ADA provider recognition program, and an evidence-based meta-analysis on new therapies. However, he made sure to emphasize (quite appropriately for a scientist!) that it is also necessary for people to continue to question whether or not these results are the right ones.
  • Mathematicians have done a “tremendous job” helping physicians. Dr. Rizza noted that mathematicians have developed a health informatics system to assess patterns of care and patient outcomes. They have also introduced telecommunication, which has, in turn, increased patient education and empowerment – one of “the most important things.”


Childhood Diabetes: What is the Way Forward?

Robert Slover, MD (Barbara Davis Center, Aurora, CO)

In this talk, Dr. Slover characterized the most pressing challenges in the pediatric diabetes care arena. He mentioned that conferences such as Keystone that are well attended by experts in both pediatric and adult diabetes represent an excellent opportunity to gain a more comprehensive understanding of the time course of diabetes as well as the related cost and care considerations. While there is consensus on the importance of involved care, patient education, social support, and novel technology, Dr. Slover also acknowledged that implementing and paying for these measures is challenging. As with many other Keystone speakers from the Barbara Davis Center, he extolled the virtues of telehealth and shared medical appointments – in this cost focused environment, we surely see both as promising alternatives to care delivery. He also emphasized the need for more mid-level providers given the shortage of physicians in the field. We were glad to hear Dr. Slover call for aggressive lobbying for improved reimbursement for time spent supporting patients by telephone and email – this strikes us as a significant challenge, though one that must clearly be overcome if diabetes care is to improve measurably and at a reasonable cost.


Panel Discussion

Richard Insel, MD (JDRF, New York NY); Robert Ratner, MD (American Diabetes Association, Alexandria, VA); Jonathan Lord, MD (University of Miami Miller School of Medicine, Miami, FL); Robert Slover, MD (Barbara Davis Center, Aurora, CO); Georgeanna Klingensmith, MD (Barbara Davis Center, Aurora, CO)

Dr. Klingensmith: Can you discuss how we address disease management in a cost effective manner? Disease diagnosis occurs across the course of the diabetes. What is the least expensive way to manage this?

Dr. Slover: I think we need more of the primary care centered model with coordination services. The best way to describe coordination of care is as part of the bundled payments of care.

Dr. Ratner: We need to consider overall cost management. We need to go back to disease management. I would say that 15% of patients in the plan are responsible for 85% of costs. We need to start doing disease management on that 15% of patients. One critical piece of the Affordable Care Act that will be different if we follow through – and that is a big if – is the concept of paying for outcomes. You get away from fee-for-service and instead aim to keep people out of the hospital. If you look at ADA economic costs, almost 50% of cost is hospitalization. If you can keep people out of the hospital, you can save costs.

Dr. Lord: Nothing is really new. New devices and new technologies have augmented our abilities to help patients. At the end of the day, great care is great care, and great care is about engaging patients with regiments that work in context of their life and getting them to be active partners. We are going to challenge the health system with academics and research. Patients are treated as inert, but we need to treat them as an active determinant of outcomes.

Q: A lot of this conference has to do with research and academic physicians. There are some of us who are doing the day-to-day work, not in research, not in academics, that are frustrated with the minimal amount of discussion around patient responsibility for their care. You’re right that we have great technology, but most of it is not going to be an option for the general population. You all have a lot of knowledge, but most patients are not going to be seeing you. I think that it’s really important for the patients to understand that it is not the doctor or the researcher or the technology or the nurse that will take care of their diabetes: they themselves will be taking care of their diabetes. We need certified educators who understand the science and can bring the science down to the patient level. I also wanted to say that you, Dr. Ratner, were pretty tough on nurses regarding school safety. I was involved when we were trying to teach teachers and parents how to inject insulin as well as glucagon, and the teachers were not very cooperative. I felt that your presentation was a little tough on us — it wasn’t the nurses that said, “It should be me or nobody.”

Dr. Klingensmith: I don’t think anyone here meant to denigrate nurse educators. We couldn’t survive without them. The question is how we can empower patients to take over their care, and be more responsible for their care.

Dr. Slover: We ought to have licenses for people to be parents, but that’s not possible. I think we live in a time when kids feel entitled, and we live with many parents who are very reluctant to impose any sort of discipline. These are some basic societal issues. And yes, you can say that it’s not me the doctor, it’s not me the educator, it’s up to the patient. But we work best in partnership. One positive story: I saw a patient by telehealth from Riverton, Wyoming. I told him that before I sign off on a driver’s license for him, he would have to bring his A1c down to 9%, and he came down to that range. Every now and then you come across someone who is wiling to take responsibility. In terms of what this conference can do, we can have more discussion groups about how we bring the science to the clinic, but I think that those are very difficult issues, and I don’t think that Bob was really bashing school nurses.

Dr. Ratner: Let me reply to that. I am incredibly supportive of diabetes educators of all kinds. To be blunt, the case in California was taken to the Supreme Court by the school nurses union, which said that the only people who are capable of administering insulin are licensed nurses, to the point that teenagers in school were unable to administer their own insulin. We all have our turf that we’re trying to protect, and at a point we need to take a step back and ask what is best for our patients. Diabetes is such a difficult disease to live with. Can you think of another disease where we expect the patients to do the biochemical monitoring, make treatment decisions, and administer the treatment? We’re asking a lot of our patients, and we need to respect them. It’s shared decision making, we need to ask our patients where can we make improvements. We can’t expect them to be perfect from day one.

Dr. Jay Skyler (University of Miami Miller School of Medicine, Miami, FL): I’m intrigued that Dr. Ratner and Dr. Slover said one of the great things we have in cost effectiveness analysis is that metabolic control that can cut costs. Dr. Lord said that when insurance industries get involved, there is a sort of death cycle, and they cut something if they provide something else. The benefits of cost effectiveness are seen down the line, not this year, which is inconsistent with how our insurance operates with revenues on a year-to-year basis. With the patients not on the same insurance watch in the future, it may not be economically feasible for insurance companies to support long-term cost effectiveness. My question is, why can’t we advocate more a single payer system? I was thinking we should be moving much more in that direction. Bob, you have had experience dealing with Congress.

Dr. Lord: There is no perfect system. If you look at the UK, there is a perfect infrastructure, but when you have large monolithic systems, they do not necessarily become smarter but slower in the adoption of new policies and such. Their bodies grow but their brains do not. The distance between the toe and the brain gets longer and longer, and the understanding of what is going on is deferred. One question we ask is how do we do things that are manageable locally. With large monolithic single payers, you need shared decision-making and contextualized decision-making. You need to engage the individual, and you have to balance everything. The dialogue needs to move more toward the dialogue on how to engage people in the conversation about health. We need to discuss how to motivate individuals to do things they probably want to do. We need to understand that it is a combination of accountability on the part of the individual and a set of systems that works for everyone. I don’t think it is a political system but more how we work with people and how we redefine everything in light of new technologies.

Dr. Ratner: The description you are giving is an ongoing process, but the process takes time. We’re changing incentives and revenue. The Affordable Care Act precludes exclusion for preexisting conditions. Any insurance has to take you without increased payment if you have diabetes. There are also fewer companies. There are three insurers in the entire state of Minnesota. The movement is pretty mild, and the insurance is going to get them anyway. Everyone, including insurance, is starting to change their thinking to longer-term thinking. If you can keep people out of the hospital, you save money. The vertical integration of medicine is driving that. That is why hospitals are buying practices. They are going to keep growing with this revenue while keeping people out of the hospital.

Q: One of the big challenges of the ACA’s goal of achieving better outcomes at lower cost is that some of the inefficiencies in the delivery system that are due to a lack of integration. I think this is an important issue, and that integrated systems like Kaiser and Group Health should serve as models for providing more efficient care, although they’re not perfect. It would behoove the ACA to stimulate integration. Are there any financial incentives for improving integration?

Dr. Lord: When I was in the hospital business, I used to say that I could find the doctors’ parking lot by looking for “no change” bumper stickers. I think that the challenge of sharing and partnering is something that’s not going to be legislated. It’s a cultural phenomenon rather than a political phenomenon. The problem with physician acquisitions is that hospitals are clueless about managing the practices they acquire. We get stuck in a cycle where we haven’t improved things, and have just spent a lot of money, and we don’t make care better. The dialogue in communities needs to be very simple, and it needs to ask how we can make care better for groups or the population as a whole. If we did that, we’d be better off.

Dr. Slover: The national healthcare systems in Britain and Canada and Europe grew out of their own cultural identity. We live in a different culture. No matter what any of us thinks, I think we can agree that we will have to build off the culture we have. Yes, Kaiser is an excellent example. That said I would ask a question: right now, or in the next five years, what do I need to do to convince somebody that it is worthwhile to pay for all the expensive supportive therapy that diabetes involves? Because nobody seems to care.

Dr. Klingensmith: Kaiser Colorado does care; we contract with them to provide care for kids with diabetes. There are no co-pays for visits.

Comment: I work in an organization integrated into the whole of the hospital, and I see the whole hospital changing. We have 70% of patients with managed care, and two other companies have asked to contract with us. We get to choose how to spend the money. We are not governed except for a few Medicare laws. It is paid for because there is no more pay for service and or volume. It is based more on outcome. However, now you are going out and making community pediatricians medical homes. We have a lot of clinics, and that is where you see fragmented care. They have abdicated everything to endocrinology, and the endocrinologists cannot be PCPs. There is one researcher from UCLA found that, at one center, endocrinologists and PCPs had to get together and see who would give the flu shot because they had not yet chosen. The medical home – PCPs and pediatricians – have staff and get paid for each member per month. We tell them to keep their data so that if they are lost we can follow-up with them.

Comment: I do clinical research with several clinical endocrinologists. As you guys are advocating for more mid-level providers in type 2 diabetes, I would hope that you also advocate for making it easier for primary care providers and pediatricians to provide the medication their patients need. We give them great medications in endocrinology, but then the primary care doctor plays hell with keeping them on those medications and puts them back on SFUs instead.


Plenary: Real Life Diabetes Management Issues

A Specialist’s/Internist’s Perspective

Richard Bergenstal, MD (International Diabetes Center, Park Nicollet, Minneapolis, MN)

Dr. Richard Bergenstal framed his “internist perspective” on diabetes management issues around three themes: 1) systems of care are changing; 2) targets for care are changing; and 3) algorithms for care are changing. His fast-paced talk covered lots of ground, centering on the Minnesota D5 (targets for A1c, aspirin and tobacco use, blood pressure, and LDL cholesterol), macro issues in healthcare (accountable care organizations, patient-centered medical homes), and treatment algorithms. On the latter, he was most enthusiastic about the use of incretins following metformin (A1c reduction, neutral/positive effects on weight, little/no hypoglycemia), but also covered an encouraging ADA poster on the benefits of glargine added on top of metformin (much better than a sulfonylurea plus glargine or the triple combination). He did not share any specific opinions on SGLT-2 inhibitors, only noting that they are “causing a lot of discussion right now.”

  • “Systems of care are changing.” Dr. Bergenstal highlighted the discussion between population health and personalized care. On the former, he recommended reading the book Big Data to understand how things are changing on a broad scale. On the personalized side, things are moving towards diabetes healthcare teams and patient-centered care.
    • The patient-centered medical home is “very big in Minnesota now.” Dr. Bergenstal cited four reasons why: 1) reimbursement; 2) pay for performance and quality measures; 3) the patient experience; and 4) the cost of care. He noted, “we’re strongly headed down that path.” However, a recent systematic review of the patient-centered medical home (Jackson et al., Ann Intern Med 2013) concluded that current evidence is insufficient to determine its effects.
    • In discussing Accountable Care Organizations, Dr. Bergenstal highlighted results that were recently released from 32 pilot organizations across the US. His slide showed a July 16 Wall Street Journal article entitled, “Mixed results in health pilot plan.” Dr. Bergenstal highlighted that all 32 of the organizations improved quality measures, though only a few actually saved money or operated at a lower cost. Despite the mixed news coverage, he thought the data was positive overall, especially because it was only after one year.” Overall, he thinks the jury is still out on ACOs but it is worth a good effort by all parties to see if this an approach that can help us reach the much sought after “triple aim” – better care, excellent patient experience at a reasonable (most hope lower) cost.
    • “There is lots to manage with diabetes – what are the five outcome measures you would pick?” Dr. Bergenstal discussed the Minnesota D5, which established five target goals for optimal diabetes care – when it started a few years ago, these aimed for a blood pressure <130/80, LDL <100, A1c <7%, tobacco free, and taking an aspirin daily. A few measures have now been relaxed: blood pressure <140/90, an A1c <8%, and use of an aspirin if cardiovascular disease is present. Notably, the state has measures on 96% of all patients with diabetes and there is a website where every clinic is tracked and ranked. Park Nicollet has 51% of patients meeting all five targets, the highest shown on the slide; to us, the fact that only half met goal really underscores the challenges, especially at a center as well-equipped and highly regarded as Park Nicollet.
  • Targets for care are changing. Dr. Bergenstal distinguished between population/accountability/pay for performance metrics (A1c <8%, blood pressure <140/90, LDL <100) and more individualized/personalized/optimal care (A1c of 6-8%, blood pressure <140/80-85-90, LDL <100 or <70). In the spirit of the ADA/EASD position statement, Dr. Bergenstal noted, “Some people should be at an A1c of 6% and some should be at 8%.” Dr. Bergenstal also referenced the recent NEJM piece on diabetes care targets (Ali et al., 2013), noting that the number of patients with an A1c <7%, declined from 57% to 52% between 2003-2006 and 2007-2010. He wondered if this was due to ACCORD and related to more doctors personalizing and individualizing therapy (e.g., targeting a higher A1c in older patients or those with a history of severe hypoglycemia).
    • Dr. Bergenstal called for a composite target for good diabetes care: optimizing A1c, minimizing hypoglycemia (including minimizing glycemic variability), avoiding weight gain (ideally, weight loss), and improving quality of life. Regarding hypoglycemia, he highlighted a publication from this morning (!) in Diabetes (Fisher et al.), where hypoglycemic episodes in rats led to cardiac arrhythmias and death. “There is more and more data coming out to say that hypoglycemia is not a good thing.”
  • Algorithms for care are changing. Dr. Bergenstal put diabetes care approaches on a continuum, with the new AACE approach on one side (very algorithmic) and personalized care (e.g., genomics) on the other side. In his view, the 2012 ADA/EASD position statement tried to “thread the needle,” incorporating algorithmic and personalized elements from both sides of the spectrum.
    • In discussing treatment algorithms, Dr. Bergenstal was most positive on the use of incretins, highlighting their composite ability to reduce A1c, minimize hypoglycemia, and avoid weight gain (or promote weight loss). He covered two studies comparing the ability of liraglutide and exenatide once-weekly to get 40% and 48% of patients (respectively) to an A1c <7% without weight gain or hypoglycemia (Zinman et al., Diabetes Obes Metab 2012; Bergenstal et al., Diabetes Obes Metab 2013). This was noticeably higher than other therapy classes.
    • Regarding incretins and concerns over pancreatitis, “that is being sorted out.” Said Dr. Bergenstal, “We need a lot more sorting out to get that answered.”
    • Showing a slide of the main ADA/EASD position statement graphic, Dr. Bergenstal added large ‘X’ marks over the sulfonylurea and TZD boxes. He highlighted the virtues of the use of incretins or basal insulin on top of metformin. He mentioned that he looked forward to the results of the just starting NIDDK sponsored GRADE trial (a RCT evaluating which is the most appropriate agent to add after metformin in patients with type 2 diabetes not in control on metformin). Dr. Bergenstal also highlighted a poster from ADA 2013 (DeVries et al.) that compared metformin + glargine vs. a sulfonylurea + glargine vs. metformin + sulfonylurea + glargine. Notably, metformin + glargine did the best on A1c, hypoglycemia, and weight gain.


Pediatrician’s Perspective

Desmond Schatz, MD (University of Florida, Gainesville, FL)

Dr. Desmond Schatz began by saying he wanted to use this opportunity to “give a call to action” to change pediatric diabetes care because we cannot afford to do nothing. He emphasized the consistent 3-5% increase in type 1 diabetes in the US and also the worldwide epidemic, showing a predicted trend for rapid increase in the incidence of type 1 diabetes in Finnish children under 16 years. Returning to the US, Dr. Schatz noted glycemic control is reaching only a minority of pediatric patients. Another problem Dr. Schatz noted were the varying guidelines for A1c levels for different pediatric age groups, with a concern about the graded scale used by the ADA. Dr. Schatz also touched on type 2 diabetes in the later part of his talk, presenting data from the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) trial that show that the majority of children with type 2 diabetes fail treatment within the first year. Additionally, he highlighted that youth with type 2 diabetes are more likely to be minorities with lower socioeconomic status and have higher rates of depression. Dr. Schatz concluded his talk by coming back to his original point: current therapeutic approaches for children with diabetes are inadequate.

  • Treatment goals for children with diabetes are different from adult treatment goals and are not entirely glucocentric; other goals including normal growth and psychosocial adjustment. Dr. Schatz emphasized this difference from adults, and noted that children have additional factors HCPs have to take into consideration. Other goals include maintenance of blood glucose lipid levels, normal BMI, and normal blood pressure, surveillance for comorbidities and prevention of microvascular and macrovascular complications, hypoglycemia, and diabetic ketoacidosis.  He also emphasized the importance of coming up with setting realistic goals for the family.
  • Fewer than 19% of children are meeting all ADA goals. Under the age of six, 13% of children meet all ADA goals, 18% between the ages of 6 and 12, and 9% between 13 and 19. These goals include blood pressure, HDL cholesterol, fasting LDL, fasting triglycerides, and BMI.
  • Dr. Schatz provided a list of diabetes management factors that are associated with better control, measured as A1c levels ≤7% (excellent control) or ≥9& (poor control). These factors were more frequent BG monitoring, bolusing before meals, missing fewer boluses, counting carbs, more frequent pump use, and less daily insulin use. Dr. Schatz also presented both sociodemographic factors associated with better control (higher income, private insurance, younger age, being non-Hispanic white), which he emphasized were uncontrollable.
  • While the ADA has a graded guideline for A1c goals in children, the International Society for Pediatric and Adolescent Diabetes (ISPAD) has a flat goal of 7.5%. The ADA has A1c goals of <8.5% for kids <6, <8% for children between 6 and 12, and <7.5% for children between 13 and 19. Dr. Schatz discussed the difference in these measures, noting that younger children have higher A1c levels. He acknowledged by not forcing these numbers down, the ADA is preventing potentially harmful effects of hypoglycemia that may accompany lower A1c levels. However, he also noted that this graded system creates a steep decline in the percent of patients who reach goal A1c levels as patients age; one study reported 65% of children under six meeting goal, 46% of those between six and 13, and 23% of those between 13 and 19. By ISPAD guidelines, the percentage of children reaching goals remained relatively consistent around 30% according to one study (Wood JR et al. presented at 37th ISPAD, Miami, 2011).
  • There are racial differences in pediatric A1c levels, even after controlling for socioeconomic status. Black children have higher A1c levels than both non-Hispanic white children and Hispanic white children. This is consistent across age groups. Similarly, black children use pumps less frequently than non-Hispanic white children and Hispanic children across all income levels.
  • Dr. Schatz reviewed the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) trial for children with type 2 diabetes children, emphasizing that the majority of treatment failed within the first year. Metformin and rosiglitazone was more effective than metformin monotherapy, however, both were unsuccessful. Although Dr. Schatz emphasized boys and girls failed at an equal rate, metformin monotherapy and metformin and rosiglitazone were not as effective for boys as they were for girls. To add insult to injury, there are high and increasing rates of hypertension, microalbuminuria, dyslipidemia, depression, and even evidence for end organ cardiac damage. Dr. Schatz also noted that children with type 2 diabetes have a rapid deterioration of beta cell function. To read more about the results of the TODAY trial, read our coverage at


NHS – United Kingdom Perspective

Stephanie Amiel, MD (Kings College School of Medicine, London, UK)

Dr. Stephanie Amiel discussed the treatment of diabetes in the United Kingdom. The total number of diabetes patients in the UK is estimated to be just shy of 3 million (a prevalence rate of 4.5%). Dr. Amiel began by highlighting key aspects of the UK’s mammoth National Health Service, a single-payer system that covers the nation’s entire population. In the UK, diabetes care is primarily hospital specialist-based, although the NHS model is evolving to make the disease a greater primary care priority. Integrated multidisciplinary diabetes clinical services have produced impressive reductions in diabetes-related renal and antenatal complications. Community-based eye screening services have achieved an 80% coverage rate among patients. The Quality and Outcomes Framework, a voluntary incentive program for general practitioners, introduced less than a decade ago, has led to substantial decreases in A1c values. Nonetheless, Dr. Amiel noted, the NHS faces a number of challenges in diabetes care. Resource limitations threaten to restrict the range of diabetes care devices and drugs that it can provide for patients. Also, citizens on the lower end of the socioeconomic status utilize healthcare less, and as a result have a higher prevalence of diabetes. She ended by noting that diabetes patients in the UK achieve comparable A1c levels to insured patients in the US and fare far better than uninsured US patients. Given that the UK has just one payer and thus a far better way to check metrics, we were disappointed that she did not give more precise figures; saying “comparable to A1c levels to insured patients in the US” is a very general estimate given that insured patients in the US have A1cs all over the map.


Asian Perspective – Implementation Issues

Linong Ji, MD (Peking University People’s Hospital, Beijing, China)

Dr. Linong Ji framed the subject of diabetes care in China with some unsettling statistics: the prevalence of diabetes in the Chinese adult population has skyrocketed in the past decade to almost 10%, with an estimated 60% of cases undiagnosed. He spent most of the talk discussing the China Cardiometabolic Registries’ observational, cross-sectional, multi-center “3B” study, which aimed to evaluate cardiovascular risk factors in type 2 diabetes patients. The observational, cross-sectional, multi-center study demonstrated that the majority of Chinese type 2 diabetes patients have comorbidities, including hypertension and dyslipidemia. Dr. Ji stated that achieving control of cardiovascular risk factors remains one of the biggest challenges for healthcare providers. Over half of diagnosed Chinese type 2 diabetes patients are on oral agents (metformin is the most prescribed, followed by sulfonylureas and alpha-glucosidase inhibitors) – most of the rest are on insulin. Dr. Ji noted that incretins have yet to establish a significant foothold in China. Ultimately, he said, the Chinese healthcare system needs to develop more of a capacity to deal with chronic conditions such as diabetes for there to be significant improvement in diabetes patient outcomes.


Best Use of Capitated Health Care System

José Armas, MD (CEO, Medical Care Consultants Inc., Miami, FL)

Dr. José Armas pressed that the US is moving to a single payer healthcare system. As evidence, he highlighted that under the Affordable Care Act (ACA), Medicaid could gain 20 to 25 million members. Dr. Armas also sees healthcare delivery models moving away from solo practices to full capitation entities. For background, under a capitation reimbursement model, a provider is paid a predetermined amount of money to care for a patient regardless of how much treatment that patient ends up requiring. Thus, the provider is financially incentivized to provide preventative services. Dr. Armas remarked that academic institutions have been particularly quick to move toward the capitation model, whereas hospital-based systems have been more hesitant due to concerns about incurring the financial risk (i.e., patients requiring more care than the capitation amount covers). Dr. Armas believes hospital-based systems will follow suit since the federal government is strongly pressing for reimbursement to shift toward a capitated model. In closing, Dr. Armas challenged physicians to stay ahead and adapt to these changes, so that they can provide better care to patients. 

  • For background, capitation is a reimbursement model in which a provider is paid a set amount for each of their patients (per period of time), whether or not that person seeks care. Thus, the provider is financially incentivized to prevent a patient from requiring services. Conversely, they are also financially at risk if a patient requires expensive treatment. In a full capitation system, the predetermined payment covers the entire package of services negotiated between a payer and provider. By contrast, in a partial capitation model, the payment only applies to some of the services provided by a provider or practice.
  • Dr. Armas sees the American medical system moving toward a single payer model. Indeed, in 2010, around the time the ACA was passed, government healthcare spending accounted for 45% of total healthcare spending at $1.2 trillion (an increase from 32% share in 1987). Medicaid membership could grow by 20 to 25 million under the ACA, which will cause the government’s share of healthcare spending to further increase. Dr. Armas remarked that it does not matter if the government spends more on healthcare via Medicare, Medicaid, or private partnerships – in the end it will still be a single payer.
  • Describing the movement toward a capitation model, Dr. Armas warned attendees that the future era of accountability and risk sharing will be difficult for solo practitioners. In particular, Dr. Armas warned that solo physician’s lack of economies of scale and could struggle under increased financial risk associated with capitation models. He believes that primary care is therefore moving away from these solo practices to management services organizations (MSO). An MSO is an organization that contracts with physicians or practices to provide administrative services and enables physicians to negotiate in union increasing their clout. In some cases the MSO is owned by a hospital.
  • Dr. Armas agreed with Dr. Richard Bergenstal’s (International Diabetes Center, Minneapolis, MN) earlier assertions that Affordable Care Organizations (ACOs) have mixed results (for details on these results, please see Dr. Bergenstal’s talk above).
  • Dr. Armas is CEO of Medical Care Consultants Inc. (MCCI), a primary care physician group of which Humana owns 49%. MCCI has 39 primary care facilities, 100 directly employed physicians, 2,000 affiliated physicians, and 865 affiliated physician groups. MCCI’s membership is 123,132 people, up from 45,578 in 2006. It operates in South Florida; Jacksonville, FL; San Antonio, TX; Corpus Christi, TX; El Paso, TX; Salt Lake City, UT; and Atlanta, GA. MCCI focused on increasing access to physicians by having a “one-stop shop” of PCPs and contracted specialists. According to Dr. Armas, it emphasizes preventative care and disease management.


Panel Discussion

Jay Skyler, MD (University of Miami Miller School of Medicine, Miami, FL); Richard Bergenstal, MD (International Diabetes Center at Park Nicollet, Minneapolis, MN); Desmond Schatz, MD (University of Florida, Gainesville, FL); Stephanie Amiel (Kings College School of Medicine, London, UK); Linong Ji, MD (Peking University People’s University, Beijing, China); José Armas, MD (MCCI Group, Miami, FL)

Q: I have one question in response to the last talk. If you are paying in advance for the system, it is to the advantage of the group to keep patients healthy and improve care. Is that the premise of the system?

Dr. Armas: We believe that the payment system 20 years ago where you were paid for the incidental care of the patient and when the patient was actually sick was not the best methodology. I can give you many, many empirical examples of why I think that. If you pay to incentivize physicians to maintain the well being of patients, and their financial alignment is such that they get paid for keeping them well, it goes without saying that this finance is incentive. It is a great incentive to provide preventative care and avoid morbidity and unnecessary costs.

Q: A quick follow-up to that—how is it in Minnesota?

Dr. Bergenstal: It is a work in progress. I think we started this HMO movement, and we didn’t like it – we did not like it acting as a gatekeeper. However, it seems to have come full circle. Equality measures are the first thing they are measured by, and people feel better about it. The movement is coming. There are still come unanswered questions: is the data really accurate? Am I being judged correctly? If you get comfortable with the clinical outcome measurement data then we are liking it better.

Q: A question that has come up many times for these metrics of care is whether, in order to meet your goals, you would exclude from your practice patients who clearly had greater challenges in meeting their targets, be they physical challenges, psychological challenges, or socioeconomic challenges. If your practice is more filled with those, you wouldn’t be meeting your goals.

Dr. Bergenstal: This has been a big worry in the Twin Cities. But it hasn’t happened as much as some feared, physicians sending some patients elsewhere, but there is no elsewhere. Now patients are getting assigned to you. The way the data works, if the patient doesn’t show up, and you have no data on them, they rate as an A1c of 9%. So you have a responsibility to get them in and help them improve their care or your scores will be bad and the patient’s health will suffer. We do think that this should be more risk-adjusted, that you should get credit for improvement. It’s hard to implement a strategy like that in a big database system, but I’m sure it will happen down the road.

Q: As a follow-up, in most systems, for pediatric patients and type 1 patients, these criteria do not apply.

Dr. Bergenstal: Its’ a huge problem. Being known in our area for diabetes, we get a lot more type 1 patients than other networks, and right now they’re all put in the same outcome database as those with T2D. For us to achieve good outcome measures, knowing that type 1s are usually more difficult to get to goal is a real effort on our part but one we are willing to keep working on. We’re working hard to parse the data more accurately but we’re not quite there yet. But yes, going back, pediatric patients are not included in our Minnesota Community Measures  (ages 18-75).

Dr. Skyler: I think it’s easier to treat a type 1 than type 2. At least you know what to do. With type 2 there are so many ways to go, and nobody knows what’s right.

Dr. Garg: What are the reasons for having different metrics for lower age groups in all the different organizations? I was also wondering if Stephanie would tell us what the UK does to that effect.

Dr. Schatz: The ISPAD calls for children to have A1c values of less than 7.5% at all ages. The ADA’s graded response was set up a few years ago. I think the concern of the ADA is the developing brain - and the harmful impact that hypoglycemia could have on the brain. We used to think that hypoglycemia only affected the brain in major ways, but now we are thinking there might be more subtle effects as well. Additionally, hyperglycemia may also have negative implications. Is it time to look at the guidelines again? Do we have real data? Probably not. Do we know when the clock for complications begins to tick? No. We are seeing complications very early on for type 2 diabetes, but we are not really sure of type 1 right now. Maybe it is time for ADA to revisit their recommendations. But as pediatricians we are cautious about the development of children. There is not really a lot of data to base decisions on, and the more intensive guidelines we have, the greater the risk there is.

Dr. Amiel: A1c levels are expected to be under 6.5% if they can be reached with one or two medications. However, if it is more difficult, and takes more than two medications, then we aim for under 7.5%. We apply different targets for different settings; if it is really difficult, you should give yourself leeway. Also we hope to extend what we do for pediatric patients to adults with type 1 diabetes. The overwhelming majority of children have type 1 diabetes. The new tariff to help patients support the price of diabetes is 10 times the pervious tariff, but it depends on if you reach certain standards. When we bring it back to the big network, A1c has already fallen, and things have improved. My concern is why it has become so much harder to have type 1 diabetes when you turn 17 and you don’t get the money.

Q: I will ask Dr. Linong Ji to comment.

Dr. Ji: Type 1 diabetes in China follows ISPAD. Right now, A1c levels for type 1 diabetes patients are around 9%.

Dr. Irl Hirsch (University of Washington, Seattle, WA): Before I get to my question, we used to do these shared appointments. We don’t anymore because it became economically disastrous for us, as we couldn’t get properly reimbursed. My comment is about what I think is a big concern – the topic of cherry picking. The largest, most prominent diabetes center in Seattle twenty years ago is virtually nonexistent now because of the risks, the costs, and the lack of financial advantage for seeing these patient populations. As someone who sees adults with diabetes, even though pediatricians don’t have to worry about this as much, I see eating disorders in fifty and sixty-year-olds, severe psychiatric and socioeconomic issues no different than pediatrics, and I’m going to get punished for it. I’m not going to get their A1cs down. On top of that, our metrics are far from perfect. We showed that for 14% of patients, their A1cs don’t even work, because of EPO or cirrhosis or anemia. And my final comment goes back to the ADA/EASD document about individualization. Whether it comes to lipids or whatever, if I have a patient who refuses to take a statin, I may not want to take care of that patient, because I won’t get his LDL down, and I’ll get dinged. I think the bigger issue is the geriatric Medicare-age type 1s, of which yesterday morning I saw three, and if I strive to get their A1c down below 7% or 7.5%, while they have a huge risk of hypoglycemia, I now become a bad doctor just to achieve these goals.

Dr. Bergenstal: This topic is fraught with problems, but it is here for us to fix because the big data is here. People can crunch numbers and look at metrics. People need to look at metrics. I completely agree with Dr. Hirsch, and we have got to step up and push for refining the quality metrics to include the parameters we feel most accurately reflect optimal care.  We can do much more with medical records today. We may have to insist that in the medical records these factors are entered so that they are accessible. We have to fix the metrics by which our care is measured, compared and rewarded instead of saying I won’t stand for it. For better for worse, quality measurement, reward and transparency are  here to stay.

Dr. Hirsch: Just don’t call it the sliding scale risk assessment [laughter].

Q: In your study Dr. Ji, you ranked hospitals as primary, secondary, and tertiary. Can you explain more about how hospitals are ranked?

Dr. Ji: We have three levels of hospitals, and they are all government owned. Primary care happens in community hospitals. Secondary hospitals are hospitals at the county level. Tertiary hospitals are affiliated with universities or are central hospitals. Those are the three levels.

Q: So it’s like community, county, then city?

Dr. Ji: Yes, that is right.

Q: As we look at population control of diabetes and chronic diseases, how will the roles of physicians and nurse practitioners differ in improving control and quality measures?

A: I would say it’s huge. I think nurse practitioners are huge. Physician assistants are doing a great deal of the work in our area. I won’t call it competition, but how it fits in the home.

A: What we are already seeing is where we refer to primary care as both nurse practitioner and PCPs in the state of Florida. Especially with Medicare and the increasing number of patients and growing shortage of primary care physicians, in our area nurse practitioners are treated equally to primary care physicians, and we expect more of the same in the future.   


Plenary: Diabetes Care Challenges in Children

How to Achieve Better Target A1c in Children

Desmond Schatz, MD (University of Florida, Gainesville, FL)

Dr. Desmond Schatz opened his thoughtful presentation on glycemic management in children stressed that “children with diabetes are not diabetic children. They are normal children who have diabetes.” With this framework in mind, he detailed the barriers to better glycemic control children with type 1 diabetes and their families face, and solutions providers can help implement. Dr. Schatz’s take-home message was that “[providers] are [children’s] cheerleader!” Thus, providers should praise freely and acknowledge that having type 1 diabetes is hard. Dr. Schatz championed the need to establish a relationship with the child and their parents. In particular, he pressed providers to “listen to the patient first; their parents can wait.” When discussing a child’s management, Dr. Schatz called for providers to find solutions, not just faults, and to stop blaming patients. Finally, he emphasized the importance of not limiting a child’s activities and emphasizing the need for a healthy diet not dieting.

  • Children have higher A1c targets than adults, due to increased hypoglycemia risk. Dr. Schatz explained that A1c targets are set higher for children since hypoglycemia can negatively impact the developing brain. Additionally, children might not recognize the symptoms of hypoglycemia or not effectively communicate these to their caretakers. Thus, a child younger than six’s target A1c is <8.5%. This converts to a blood glucose level is 100-180 mg/dl (110-200 mg/dl at bedtime) vs. adult’s recommended range of 70-120 mg/dl. As children become older, these targets gradually decrease.
  • While more children than adults are reaching their A1c target, children at target still tend to be in the minority, according to Dr. Schatz. Sixty-five percent of children younger than six years are at their A1c goal of <8.5%, 46% of children 6-12 years old are at their goal of an A1c <8.0%, and 23% of people 13-19 years old are at goal of <7.5% (DiMeglio et al., Diabetes 2012).
  • In the T1D Exchange, factors associated with better control are either related to management or sociodemographic in nature (p<0.0001 for all). Dr. Schatz emphasized that factors related to management are modifiable; these include 1) conducting more frequent blood glucose monitoring, 2) bolusing before meals, 3) missing less boluses, 4) counting carbohydrates, 5) using a pump, and 6) using less insulin daily. Sociodemographic factors associated with better control include 1) having a higher income, 2) having private insurance, 3) being younger, and 4) being non-Hispanic white (in press, Pediatric Diabetes 2013).
  • Dr. Schatz explained that glycemic control tends to deteriorate during adolescence for several reasons. One of these is that insulin resistance increases during puberty and adolescents have a higher prevalence of mental illnesses and eating disorders. Additionally teens tend to be ambivalent, impulsive, rebellious, and experimental. Dr. Schatz also highlighted the prevalence of mental health symptoms in adolescents with type 1 diabetes (n=150): 11% screened positive for depression, 21% for anxiety, and 21% for disordered eating. Adolescents with a positive screen had double the odds of having an A1c ≥8.5% (Bernstein et al., Clin Ped 2012). In a later session, Dr. Timothy Wysocki (Nemours Children’s Health System, Jacksonville, FL) noted that in many studies, depression is more prevalent than anxiety and eating disorders. Attendees expressed great uncertainty on how to treat anxiety in particular, despite realizing that it was a very big barrier.     
  • Dr. Schatz negotiates realistic blood glucose levels, weight, and behavioral goals with his adolescent patients to help improve their control. He warned attendees to avoid a perfectionist approach. He also pressed the importance of providers allowing patients to express negative feelings about having diabetes and to refer the patient to a mental health specialist if depression is suspected. Additionally, Dr. Schatz stated that providers should listen to patients’ weight and body shape concerns and that they should emphasize healthy eating not dieting.
  • To minimize risk of hypoglycemia while improving glycemic control, Dr. Schatz recommended the use of flexible insulin regimens with a pump. He also pressed that providers need to discuss hypoglycemia at each visit and educate patients on exercise and bedtime strategies, as well as providing blood glucose awareness training and information on alcohol’s effects.
  • Dr. Schatz recommended having patients use the app GOMEALS for help with meal planning. GOMEALS keeps track of meal history, has an activity tracker to calculate calories, logs and tracks blood glucose levels, generates reports and graphs, and uses the cloud to sync across multiple devices. GOMEALS is available for the iPhone, iPod Touch, iPad, and Android.
  • Dr. Schatz explained that his pediatric patients are more likely to take blood glucose readings and bolus accordingly if they understand why the measurement is needed and what action they should take. He teaches his patients that to estimate a bolus and correction dose based on the following five questions: 1) What is my blood glucose now?, 2) How many carbohydrates am I going to eat?, 3) What is my insulin dose for this amount of carbohydrate?, 4) Should I lower the dose because I plan to be very active?, and 5) What has happened previously under these circumstances?     
  • Dr. Schatz told attendees to build a relationship with their pediatric patients, and to listen to their children before the parents. Doing so requires that providers spend time with their patients. Dr. Schatz explained that he thinks it takes 45 minutes to an hour to learn a child’s unique personality and priorities, and to help brainstorm management strategies with the patient.


Issues with Type 1 Diabetes

Tim Wysocki, PhD, ABPP (Nemours Children’s Health System, Jacksonville, FL)

Dr. Tim Wysocki detailed how a teenager with type 1 diabetes’ family environment can impact their glycemic control, and how interventions can successfully address potential problems. Dr. Wysocki opened by highlighting the “striking uniformity” across study samples in how adherence declines as children age. Dr. Wysocki is “fairly well convinced” that the family environment leads directly to diabetes outcomes and can positively or negatively impact an adolescent’s adherence. Family characteristics that influence type 1 diabetes management among adolescents, according to Dr. Wysocki, include the number of adult caregivers in the home and the number of parents involved in the diabetes management, communication and conflict negotiation, parental withdrawal versus involvement, parental monitoring and adolescent disclosure, and absence versus clear expectations for self care. Notably, while a parent’s diabetes problem solving skills (ability to prevent and treat hypoglycemia and hyperglycemia) is associated with a child’s glycemic control, the child’s own diabetes problem solving skills are not. To explain this, Dr. Wysocki hypothesized that many adolescents know how to manage their diabetes better and are choosing not to. He believes this could be due to adolescents choosing to fit in with their friends without diabetes over their own glucose self management. Dr. Wysocki did not detail successful interventions; however, he noted that commonalities of effective interventions for teenagers with type 1 diabetes are that they are theoretically grounded, target diabetes-specific behavior, utilize experiential learning, are flexible and adaptive, are sensitive to cultural diversity, and utilize a sustainable delivery system.

  • Dr. Wysocki outlined five “basic behavioral principles”: 1) behavior can be strengthened either by positive reinforcement or by avoidance of adverse events, 2) behavior is more strongly controlled by immediate consequences rather than by delayed consequences (as seen with overeating, smoking, gambling, etc.), 3) behavior that is reinforced intermittently is stronger than behavior that is reinforced consistently (for example, if at a slot machine, a consistent dozen wins are followed by a loss, a person is more likely to stop than if every three times they lose), 4) avoidance of adverse events may encourage “superstitious” behaviors (Dr. Wysocki noted that superstitions can develop around avoiding hypoglycemia), and 5) uncontrollable adverse events may lead to “learned helplessness.” Learned helplessness is when a person believes it is not possible to remove themself from an unwanted situation, because it previously was not possible (though the situation may have since changed).


Panel Discussion

Robert Slover, MD (Barbara Davis Center, Aurora, CO); Desmond Schatz, MD (University of Florida, Gainesville, FL); David Dunger, MD (University of Cambridge, Cambridge, UK); Tim Wysocki, PhD, ABPP (Center for Pediatric Psychology, Jacksonville, FL); Georgeanna Klingensmith, MD (Barbara Davis Center, Aurora, CO)

Q: We have a lot of data, we have a lot of great ideas, and an increasing number of tools. But when we go back to our clinics on Monday, we’ll again face poor reimbursement and too many patients. I’d like to ask the panel about things we can stop doing to make time for the exciting things we could be doing.

Dr. Schatz: On a certain level, there are things we have to do for reimbursement, such as all the documentation. I can’t give you a practical solution. We need to change behavior; we have to educate the insurance companies and others about what is really necessary.

Dr. Dunger: Can I make a controversial suggestion? I have noticed in my clinic and others in the UK that nurses and doctors are spending an increasing amount of time putting people onto pumps and sensors who have an A1c under 7.5%. And the other 50% of patients are languishing; they are doing so badly without our help. I think we need to spend more time helping these people rather than tinkering with an A1c <7.5%.

Dr. Klingensmith: The other controversial question is whether you can really do anything with patients over 10%. Are there too many psychosocial issues? That type of issue may require psychiatric care.

Dr. Slover: I can add to that. As physicians, our model has always been an educational model and we have learned today that that just is not enough. We see over and over against the notion of “if I could just explain this better or help them understand the consequence.” I am wondering if part of what we need to do is strengthen their social support. Maybe we need to go back to our medical school education and improve the training on how we support patients with chronic conditions.

Dr. Wysocki: the Affordable Care Act includes inducements to establish just that. It’s pretty clear that diabetes is one of several pediatric chronic conditions, asthma another, that require this type of model. But in answer to your question, Marian, I think you’ve got to find the number of words available to you during a visit, and choose them wisely, to focus heavily on encouragement and praise and catching kids being good, rather than catastrophizing, lecturing, those kind of “guilt trip” communications which alienate a lot of these kids on the high end of the A1c spectrum.

Dr. Francine Kaufman (Medtronic Diabetes, Northridge CA): I think we are misdiagnosing our patients. I think these children have social cancer. Where I am, 10% of my clinic’s patients are homeless and their parents are addicted to drugs. Many, many years ago in the ‘80s, we had a residential treatment center for about a year and a half, for heavy hitters who had social cancer. They all did extremely well in the residential center but when we returned them home, if it had not been corrected, they started doing very poorly again. I think that we need to work with other people to help us on these issues.

Q: The talks mentioned the complexities of MDI [multiple daily injections] and high A1c. Is there anything you know from literature about using NPH insulin or premixed 70/30, and maybe that takes the complexity away?

Dr. Klingensmith: We certainly have done and do that. It is not a way to actually improve the A1c dramatically. It makes the patients feel like you have heard their concerns and issues. We do occasionally have kids who are moved from a simple regimen to an MDI regimen, and they forget to take their insulin. I can’t say that I have seen dramatic decreases when you revert to a BID [twice daily] 70/30. Lori Laffel has a report of giving rapid acting NPH and Lantus at noon at school and she did get a decrease in A1c with that.

Q: The question is how to achieve a better target A1c. What we have not discussed at all is furthering a person’s endogenous insulin production. I think that good control breeds good control. A person who does well the first few years will continue doing five SMBG a day, whereas s a person who has a difficult start will not. 

Dr. Schatz: We are trying. We are trying different agents. I don't think we really know the outcomes beyond two years. There certainly are a number of studies that showed what I would say is preserved C-peptide in the first six months that is then taken out at two or three years. We have some very preliminary data using a low dose ATG-GCSF. I think we are making progress. Have we been able to halt the fall in C-peptide? The answer is no. Again, we are trying. It remains to be seen if the improvements in C-peptide we see at two years are going to lead to better outcomes, just as the improvement in DCCT certainly led to a decrease in complications later on.


Session: Pediatrics

Update on Type 2 Diabetes in Children

Philip Zeitler, MD (University of Colorado, Aurora, CO)

Dr. Philip Zeitler’s spent most of his time discussing the results of the Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) trial, which was published in the NEJM on May 3 of 2010 (read our full coverage of the study results at The study aimed to investigate the efficacy of metformin alone, metformin and lifestyle modification, or metformin and rosiglitazone (GSK’s Avandia) in improving pediatric type 2 patients’ A1c levels. The rosiglitazone-metformin co-therapy group performed significantly better than the metformin monotherapy group (in which therapy failed for over half of the patients by the end of the trial). Dr. Zeitler highlighted significant efficacy differences between the groups based on gender and race. For example, lifestyle intervention was far more effective in boys, while rosiglitazone performed better in girls. This variation, he noted, underscores the importance of individualizing therapies for different patients. A1c was found to be the primary predictor of success (or failure) with metformin. Overall, the results of the TODAY study indicate that metformin monotherapy is inadequate for a significant number of pediatric type 2 patients, potentially indicating a need to expand treatment options for this small but rapidly growing patient population.


Panel Discussion

Paul Wadwa, MD (Barbara Davis Center, Aurora, CO); Andrea Steck, MD (Barbara Davis Center, Aurora, CO); Marian Rewers, MD, PhD (Barbara Davis Center, Aurora, CO); Jennifer Barker, MD (Children’s Hospital, Aurora, CO); Philip Zeitler (University of Denver, Aurora, CO)

Q: My question is related to screening for thyroid issues for type 1. I’m a general pediatrician, so I generally use a TSH cascade, and I’m curious about the value of doing a TPL automatically.

Dr. Barker: I think the TPL will identify people who are at a higher risk for developing thyroid disease in people with or without type 1. But there are some people who show up negative on the TPL who go on to develop the disease. I would say that the TSH is generally enough.

Dr. Rewers: In type 1 diabetes, I screen everyone for both. The TPO may guide the frequency of the TSH test. If the kid is negative, you can probably screen every two years rather than every year.

Dr. Barker: The other thing I would say is that if they had a positive TPL, you would treat them earlier.

Dr. Zeitler: The difference here may be the difference in your background population. If you’re screening a background population, a TPL may not be useful, but in a type 1 population it may be useful.

Q:  Should we screen for Addison’s Disease in people with type 1 diabetes?

Dr. Barker: There are no recommendations for that. If you have a reliable assay, I think that it is reasonable to screen for it once at diagnosis, especially if there are symptoms. I think it depends on the assay. If you have an unreliable assay, I would not, because you are not going to trust the assay.

Dr. Zeitler: I don't have an opinion in type 1 diabetes.

Dr. Rewers: I would screen. It is an inexpensive screen to do every year or every two years. Even screening every five years would probably prevent many of the events.

Dr. Steck: I would screen at diagnosis and then with symptoms.

Dr. Wadwa: I would screen as well at diagnosis. I am not sure if I would screen annually thereafter.

Q: For the kids who have prediabetes, how many of those will go on to get type 2 diabetes, and will lifestyle interventions make a difference for them? Do you have any suggestions on how to help them be successful?

A: This is only an hour-long talk — the short version is that we don’t know enough about prediabetes in adolescents to know. We do know that there is a very high remission rate with no intervention, especially after puberty, because of the reduction in insulin resistance. There hasn’t been a study of interventions to prevent diabetes in kids with prediabetes. The progression from prediabetes to type 2 is actually quite low in kids with IFG, but it’s a common finding in kids with IGT. The kids who progress are usually those with combined IFG and IGT. That may be the area where intervention is most needed. In terms of what works, I honestly don’t know. Prayer? We know that some things work: diet, weight loss, decreased carbohydrate intake, increased activity. What we don’t know is how to deliver these things to challenging populations, and I would be making up any tips I was to give you now.

Dr. Wadwa: Stay tuned, next year we might have some more information.


Meet the Peers – Pediatric

Pediatric Providers

Jennifer Barker, MD (Barbara Davis Center for Diabetes, Aurora, CO); H. Peter Chase, MD (Barbara Davis Center for Diabetes, Aurora, CO); Georgeanna Klingensmith, MD (Barbara Davis Center for Diabetes, Aurora, CO); David Maahs, MD (Barbara Davis Center for Diabetes, Aurora, CO); Marian Rewers, MD (Barbara Davis Center for Diabetes, Aurora, CO); Paul Wadwa, MD (Barbara Davis Center for Diabetes, Aurora, CO); Philip Zeitler, MD (Children’s Hospital Colorado, Aurora, CO); Tim Wysocki, PhD, ABPP (Nemours Children’s Health System, Jacksonville, FL)

Q: When it comes to introducing solid food, four months seems to be a magical time. I don't understand why risk goes up with introduction after six months.

Dr. Rewers: It is a great question. There is a well described epidemic of celiac disease in Sweden, which was triggered not by getting wheat products or solid foods too soon but was caused by getting a lot of wheat based products later on. So basically the Swedish academia of pediatrics decided to make a recommendation not to introduce any wheat based products until after six months. But the industry picked up on this and started making a lot of wheat based products for that age. So kids went from being wheat naïve to having a large amount of wheat products, and they were not being breast fed at that time.

Q: Have they thought about biopsying the pancreases of new onset patients who have died?

Dr. Rewers: It would not pass our IRBs in the US. There was a group doing it in Japan – the results were controversial. A group in Norway is also looking into this. They were investigating a drug, and the intervention failed but in the context of this study people consented to have a pancreas biopsy performed. It was pretty big sample, I think a cubic centimeter. There is an exciting paper coming out showing that compared to similar samples obtained from controls; five out of six of those patients with type 1 diabetes had the presence of an enterovirus in their pancreas. Notably there was staining for the enzyme the virus makes to propagate itself. I think to date this is the strongest piece of evidence that there might be a sustained viral infection in people with diabetes. Is it the cause or a result of them having type 1 diabetes? It is not clear. It has also been found in the thyroid samples, so it might bot be specific to the pancreas.

Q: I have a child who is vomiting who has mild to moderate ketones and blood sugars in the 80s. I am having a hard time keeping him up. Is the use of glucagon ok?

Dr. Klingensmith: I think there are a couple of reports of using low-dose glucagon. We use that frequently.

Q: Even with the vomiting?

Dr. Klingensmith: Yes because it is very low dose. Not as low as the bi-hormonal pump but it is a very low dose.

Q: Would you use zofram in a kid who is vomiting?

Comment: It is very expensive.

Dr. Chase: I think that zofram has prevented more ER visits than almost any other drug.

Q: Do you make sure that the family has a ketone meter, before prescribing zofram?

Comment: I mean that is the same reason I would never use a low-glucose suspend pump because they might go into DKA [laughter]; isn’t that what the FDA says? We try to make sure that every family has a urine ketone meter. We also give out some blood ketone meters especially to kids on pumps but those are expensive so some people definitely might not have that.

Comment: I am a PCP. If the Barbara Davis Center goes and sends all of my patients into the ER, my Accountable Care Organization is going to ding me because my patients are going into the ER too often. I think we have too look at why are they sick. In some situations use the zofram. If they have too many ketones or they have a rigid belly, send them to the ER.

Q: I have a pre-adolescent who is autoantibody negative. Do you mange that kid differently? At what point do you look for MODY?

Dr. Klingensmith: I think some people say that a child who is under 10 years old, is slender and non-Hispanic white might not need to have antibodies tested, unless people like the BDC can do that. Also take a good family history, although we now think that family history is negative half the time for MODY – they are new mutations. MODY testing is expensive, so if you are talking about the cost of care you probably want to look at C-peptide after two or three years.

Comment: The test is a lot cheaper than the cost of insulin for those years.

Dr. Klingensmith: That is true, except you have to make that case to the insurance company.

Dr. Barker: Also how many kids do you have to test before you find that one kid with MODY?

Q: Haven’t there been reports that 60% of people with type 2 diabetes have autoantibodies?

Dr. Chase: About 10-14% of people with type 2 diabetes have autoantibodies, most commonly the GAD autoantibody. If you consider that and then look at the number of people who develop type 1 diabetes, there are actually more adults now developing type 1 diabetes in a given year than there are children getting type 1 diabetes.

Dr. Maahs: I think it is a very important issue. Adults who have type 1 diabetes are often diagnosed with type 2 just because of their age. We have a pediatric patient who was autoantibody positive. We knew that the mom was also autoantibody positive. She went to the hospital and what happened? She was diagnosed with type 2 diabetes. Most cases of type 1 diabetes happen after the age of 18. Most people with type 1 diabetes are diagnosed as adults.

Q: What are you going to do with type 2 patients given the TODAY data? Are you testing other agents? The whole thing with the pancreas, is making me wonder about the GLP-1s. I had had hope to use those drugs in my kids.

Dr. Zeitler: There is obviously interest in other agents. The TZDs are probably dead.

Q: Even though Avandia has been brought back?

Dr. Zeitler: It has sort of been brought back. Nobody is going to prescribe it. All of the drug companies are interested in doing studies in pediatric patients with type 2 diabetes. They are all going to fail for futility. They are not enrolling well. Although there is great interest I don't see approvals coming for any new indications anytime soon.

Q: Are you going straight to insulin?

Dr. Zeitler: Yes.

Q: My patients are all coming from problematic families. What insulin do you use with them?

Dr. Zeitler: I use basal insulin. That will do the trick for a very long time.

Q: Have you been measuring your postprandial blood glucose?

Dr. Zeitler: No, because they do not measure their blood glucose.

Comment: I have measured their postprandials, and they usually are high.

Dr. Zeitler: I am sure they are.

Comment: I think a lot of these kids need to have boluses.

Dr. Zeitler: Do they take it?

Comment: Some.

Dr. Zeitler: I would rather have them do what they will do, rather than tell them to do the ideal, have them not do that, and not come back. Giving them something they can’t do causes them to feel bad and not come back.

Dr. Wadwa: I worry about the risk of hypoglycemia in those situations with boluses if they are not doing it right. While with what Phil is suggesting that risk is a lot lower.

Comment: This is all anecdotal. Some of my patients will do it and do better.

Dr. Zeitler: Some will do that.

Dr. Klingensmith: Des showed some information for screening for depression, anxiety, and eating disorders. Depression was present in 10% but anxiety and eating disorders were present in 20%. Tim, what do you think about that?

Dr. Wysocki: That data runs counter to the data you tend to see where depression is the most frequent followed by anxiety and eating disorders. We can bicker about the order but when they are present you are in trouble. It would behoove us to screen for them.

Dr. Klingensmith: Screen everyone?

Dr. Wysocki: In terms of eating disorders in diabetes, there is evidence that it occurs very rarely in males. If I am seeing a female who is losing weight and maintaining a very poor A1c, I would probably look into an eating disorder. I think that depression screening makes sense universally in adolescence and then to do that screening more regularly with symptoms thereafter.

Comment: Most of us have some sense of how to treat depression but anxiety probably makes us all anxious.

Dr. Klingensmith: And it is a terrible barrier.

Comment: It is, and I don't think we know what to do.

Dr. Wysocki: The data in kids without diabetes shows that medications and behavioral therapy are about equally effective. There are a lot more cognitive therapists than there are child psychiatrists. The child psychiatrists in my group are very nervous about giving another medication to a kid with a chronic disease, which counters their reputation. The best prognosis for cognitive behavioral therapy is with a more specific phobia; when you start to get into more generalized anxiety, medications become far more important.

Q: What are your thoughts about adding metformin to kids with type 1 diabetes who are overweight and are struggling on high insulin levels.

Dr. Klingensmith: I am all for it.

Dr. Chase: It would be nice to have more data on that.


Meet the Peers – Adult

Adult Educators

Mary Voelmle, NP (Barbara Davis Center, Aurora, CO); Christie Beatson, MS, RD,CDE (Barbara Davis Center, Aurora, CO); Mindy Schwartz, RN, MS, CDE (Marin General Hospital, Greenbrae, CA); Terra Thompson, BSN (Barbara Davis Center, Aurora, CO); Sonya Walker, RN, MBA (Barbara Davis Center, Aurora, CO); Lisa Meyers, MSW, CDE (Barbara Davis Center, Aurora, CO)

The informal “meet the peers” session of adult educators began with the panel asking the attendees to share their experiences and pose questions to the group. The educators attending the session came from a variety of backgrounds, including the Barbara Davis Center in Aurora, Colorado and independent centers in southern California. Regardless of geographic location, most educators agreed that the majority of their patients are on Medicare, which covers two hours of education per year after the initial 10 hours of training in the first year (not enough, in our opinion). Turning to the financing of education, some attendees express concern about maintaining their centers in an era of tightening budgets, while others shared that their affiliated providers absorbed the cost of education in the hopes that prevention would reduce long-term costs. The themes of the conversation included shared appointments, telemedicine, the transition from pediatric to adult care, ad reimbursement– a topic we have heard frequently at the conference.

  • Patient safety and happiness was key in the education process. The educators were dedicated to empowering patients to improve their health. One educator emphasized the power of positivity, specifically referencing the work of Martin Seligman, the father of positive psychology. Learn more from his website,

Questions and Answers

Q: I’ll start us off. What are your experiences with shared appointments or group appointments? We have tried them in teen programs, and some teenagers are really happy with it while some teenagers cannot seem to move to adult care fast enough so they do not have to do the group project. Has anyone else tried these programs? What are your experiences with reimbursement and copays?

A: Our center has invested in group appointments for all pregnant women – not just those with gestational diabetes. We know that we get a lot done, and we know it works if you have groups and are optimizing time and mental health. The problem is that it is hard to recruit these women and to maintain their attendance; we might start with six women the first week and then only have two the second week. It’s terrible with diabetes. It is a big problem, whether or not patients think they know everything, because they do not want to discuss it. Financially, it’s a nightmare. It’s so hard to sell.

A: We do something a little differently to accomplish some of the same goals. We have a gestational diabetes class that we used to hold weekly but now hold twice a month. We call this a medical appointment/class. They learn how to use the glucose meter as a group. We [NPs] do diabetes education in small groups, and the dietician does the second half of the class. The education is based on the need of the class. We have a follow-up after seven to 10 days. We have seen real success with the women who learn how to use the meter. During our sessions, patients do not see a primary provider, but they do see a dietician. Most women who come to us are referrals from their obstetrics.

Q: How does the shared provider appointment work? Are the patients uncomfortable talking?

A: All of our patients sign a HIPAA form. Even with that, however, they say they are still uncomfortable. We have nine clinics that have this system. Maybe if we had only one clinic we would be able to concentrate on it and really unpack why are patients do not like the program. I am for these – I think it’s a great way to provide all the services. What I think it comes down to is a behavior change for patients. They have always just seen a primary provider. Maybe patients need to see the primary provider for just a few minutes to ensure that they are getting the important parts of their visit with the primary provider. 

Comment: On the pediatric side, we have these groups for kids and teenagers. In the teenager group, we provide hypoglycemia education around college, sex, drinking… topics that patients might benefit from having other peers to talk to. After this conversation, each patient sees the provider individually.

Comment: The idea there is peer support. There is not necessarily that shared commonality among adults.

Comment: I work in a Hispanic population with low literacy, and when we tried shared appointments, everyone was terrified to talk. They were all afraid that everyone would know everything about them. They liked that all of the resources were in the same place, but they felt that their privacy was being violated. However, a facility in Kansas City has tried shared appointments, and it has been very successful in the outpatient setting.

Comment: I think we need to address what is different between a group class and a group appointment.

Comment: I think another issue is that some of the classes are at full capacity while others are nearly empty – we were discussing this earlier, and I think it is a really important point.

Comment: Another idea is to do part of the appointment as individuals and part of it as a group. Many of the patients were uncomfortable with the foot exam – the problem was the patients didn’t want other people to know they have foot problems.

Comment: Our method is similar. We have a class and then individual appointments. It works well if everyone comes. The fact that appointments are long is also a problem for efficacy – patients do not want to come to four-hour appointments.

Comment: At our facility, we also had to review basic diabetes education in the first shared session, such as what the A1c levels meant. Patients have such different needs that it is hard to do shared appointments [nods].

Comment: The other problem with shared appointment is that you kind of get one person who wants to do all of the talking, and it becomes really hard to include everyone in the conversation; you have to work on breaking down the barriers and think about who you are asking questions to. With the ADA certification, we have one initial visit with a nurse educator where we go through common education and the foot exam, and then patients have an hour with the dietitian, and then eight hours of group education. Finally, they have two hours to go to the dietician. This works well if the patients come, but everyone has an excuse. That is what we find hard. We have six to eight patients scheduled, and then either everyone comes with their spouse, or no one comes.

Comment: It is really important for any sort of group visit to create an environment of safety and confidentiality. At the beginning we set up ground rules of the meeting. The patients feel uncomfortable without ground rules because they do not know what to expect. We are changing the expected physician’s procedure. Even doing ice-breakers helps because it makes the patients feel safer. I also think that shared appointments are harder in a small community where everyone already knows everyone else.

Comment: One thing that has worked really well for us is to start off the group session with a number of questions. My staff will start asking questions, and soon everyone gets involved. When patients answer questions incorrectly, my staff will just chime in with the right answer – always staying positive.

Comment: We do three classes, each 3.5 hours. At the beginning of every class, each patient has an individual check-in with either a nurse or a dietician. At that time we say to them, is there anything you want to tell us confidentially? We also remind them that they can always do that. Also, each time they come back, we fill out a “since I saw you last” form, and that is when we ask about other exams (e.g. eye). It works well as a balance between confidentiality and sharing. Finally, addressing the issue of talkative people, I find that sometimes they are less talkative in the group than you might expect – although not always. 

Q: What can people tell us about telemedicine? What is the reimbursement like for that?

A: We started doing telemedicine about a year ago. Every month, a pediatric endocrinologist calls in to visit with patients. It saves the patients about 180 miles of driving to Denver, and it really helps if the weather is bad. At the visit, the nurses will measure A1c and vitals, download the pumps, and the patient will have a private videoconference with the physician. Afterwards, we do the after-visit summary. This allows us to go over teaching points or schedule education. Physicians cannot see the infusion set very well, so we assess those, as well. We see up to ten patients in a day, with each patient taking about 45 minutes. We only get reimbursed if there is education, which is challenging because it pulls us away from our paying positions for a whole day. We probably lose a great deal of money from this telemedicine. The reason we do not get reimbursed is because the primary providers are not affiliated with our system, they are only using our technology. We could charge a facility fee for that, but we are not going to deny anyone service if they cannot pay. So then you get into an ethical issue of asking if it’s okay to make some patients pay but not others. Also, if it is not cost saving to the patients, they would just go down to Denver. Even though we do not get reimbursed for our services, our facility is still supportive of it. I’m more pessimistic about it with the budget changes, however. We have thought about trying to do it with fewer patients, but we help a lot of patients and there is a lot to do.

Also, quite a few of our providers that do telemedicine do it in Jackson. When we use their facilities, they also have trouble with the reimbursement. That is probably why we keep doing it with Denver: because we can be on the other side sometimes.

Q: Has the telemedicine brought new patients?

A: It has helped create liaisons, as well as letting people know that they can use as a resource in the area. So it is good in that sense. 

Comment: We have been an ADA accredited center for 10 years. We have been told that nursing educators cannot charge for their services. Therefore, we do not charge for anything except for the classes.

Comment: We have a diabetes center in our endocrinology building. They are separate offices through the hospital, and the diabetes center can charge for the service the same day the doctor does. We have separate billing.

Comment: As we move into the new model with shared savings, we have protocols where I do blood sugars, education, and go over adjustment. We have found that this saves the time and money of the physicians. They have no interest in revenue from our practice because they think that our services will save money in the long run by preventing complications and hospitalizations.

Comment: Our center had to clarify what we were charging, and we had to personally make sure the trail was correct. We had to be very clear. The AADE is a great resource, and you can call them and ask how you should charge a service. The coding of these services has to be on your radar. I feel much more comfortable knowing the coding, knowing how much we charge, and how much we get reimbursed. It also solidifies the trail.

Q: Do pediatric patients want to go to the adult practice? Does it vary depending on the patient? Does it happen that physicians do not want to let their patients go?

A: When the patient has a baby, we push them to the adult clinic. If a patient’s A1c levels are 12%, when they hit 18 they are out. Sometimes the child has been going to their doctor since they were four years old. They have such a bond with their physician, and they have a hard time leaving. They may also be scared that they are going into a world where their hand is no longer held. Also, physicians start looking at their feet.

Comment: Some physicians feel that it is not appropriate to transition their patient from pediatrics to adult right as they enter college, because the patients may get lost in transition [nods of agreement].

Comment: It can really vary depending on the situation of the patient. If the patient is staying in Colorado, they can have an easy transition from their pediatric provider to an adult provider. Other kids go off to college, and they may get lost because they never set up an appointment with a physician in either their college town or their hometown. Pediatric patients in Colorado that go to the Barbara Davis Center as a child have to go to Kaiser as an adult. Many patients do not know how to get into the Kaiser system. We did a study a few years ago to see how many kids were falling through the cracks, and we found that it was a real problem. Now we have a specific adult provider we give the pediatric patients to. 

Comment: The other issue is that pediatric patients are on Medicaid, and when they hit 18 they are suddenly off.

-- by Adam Brown, Hannah Deming, Hannah Martin, Manu Venkat, and Kelly Close